Skip to content

Use Claude Code, Codex, VSCode, Pi, and OpenCode (and 6 other harnesses) for free (1.3B+ free tokens) from your terminal, app, IDE, or phone, and now from the browser with native browser sessions (multi-harness + multi-model) like OpenClaw (voice supported + ToS friendly)

Python

Retire GitHub Models and migrate selections to default (#1668)

Ali Khokharec6fda36f3e9
diff --git a/.env.example b/.env.example
index ac7562bf..14a08fde 100644
--- a/.env.example
+++ b/.env.example
@@ -60,10 +60,6 @@ BEDROCK_BASE_URL="https://bedrock-mantle.us-east-1.api.aws/v1"
# COHERE_API_KEY=<your-token>
-# GitHub Models Config (OpenAI-compatible Chat Completions at models.github.ai/inference)
-# GITHUB_MODELS_TOKEN=<your-token>
-
-
# Z.ai Config (one key for Coding Plan zai/... and pay-as-you-go zai_api/...)
# The model prefix selects api.z.ai/api/coding/paas/v4 or api.z.ai/api/paas/v4.
# ZAI_API_KEY=<your-token>
@@ -196,7 +192,7 @@ OLLAMA_BASE_URL="http://localhost:11434"
# Default model and optional Claude tier overrides
# Format: provider_type/model/name
-# Valid providers: "nvidia_nim" | "openai" | "azure_openai" | "open_router" | "gemini" | "vertex" | "deepseek" | "mistral" | "mistral_codestral" | "opencode_zen" | "opencode_go" | "vercel" | "bedrock" | "huggingface" | "cohere" | "github_models" | "wafer" | "kimi" | "kimi_code" | "minimax" | "cerebras" | "groq" | "cline_pass" | "xai" | "qwencloud" | "qwencloud_coding" | "together" | "deepinfra" | "siliconflow" | "nebius" | "chutes" | "featherless" | "agnes" | "zenmux" | "wandb" | "sambanova" | "kilo" | "fireworks" | "novita" | "cloudflare" | "zai" | "ollama_cloud" | "lmstudio" | "llamacpp" | "ollama" | "tokenrouter" | "nararoute" | "poolside" | "llm7"
+# Valid providers: "nvidia_nim" | "openai" | "azure_openai" | "open_router" | "gemini" | "vertex" | "deepseek" | "mistral" | "mistral_codestral" | "opencode_zen" | "opencode_go" | "vercel" | "bedrock" | "huggingface" | "cohere" | "wafer" | "kimi" | "kimi_code" | "minimax" | "cerebras" | "groq" | "cline_pass" | "xai" | "qwencloud" | "qwencloud_coding" | "together" | "deepinfra" | "siliconflow" | "nebius" | "chutes" | "featherless" | "agnes" | "zenmux" | "wandb" | "sambanova" | "kilo" | "fireworks" | "novita" | "cloudflare" | "zai" | "ollama_cloud" | "lmstudio" | "llamacpp" | "ollama" | "tokenrouter" | "nararoute" | "poolside" | "llm7"
# MODEL_FABLE=<provider/model-id>
# MODEL_OPUS=<provider/model-id>
# MODEL_SONNET=<provider/model-id>
@@ -232,7 +228,6 @@ MODEL="nvidia_nim/nvidia/nemotron-3-super-120b-a12b"
# FCC_SMOKE_MODEL_BEDROCK=<provider/model-id>
# FCC_SMOKE_MODEL_HUGGINGFACE=<provider/model-id>
# FCC_SMOKE_MODEL_COHERE=<provider/model-id>
-# FCC_SMOKE_MODEL_GITHUB_MODELS=<provider/model-id>
# FCC_SMOKE_MODEL_ZAI=<provider/model-id>
# FCC_SMOKE_MODEL_ZAI_API=<provider/model-id>
# FCC_SMOKE_MODEL_TOKENROUTER=<your-token>
@@ -309,7 +304,6 @@ REASONING_HAIKU=inherit
# BEDROCK_PROXY=<http-or-socks-url>
# HUGGINGFACE_PROXY=<http-or-socks-url>
# COHERE_PROXY=<http-or-socks-url>
-# GITHUB_MODELS_PROXY=<http-or-socks-url>
# ZAI_PROXY=<http-or-socks-url>
# ZAI_API_PROXY=<http-or-socks-url>
# TOKENROUTER_PROXY=<http-or-socks-url>
diff --git a/ARCHITECTURE.md b/ARCHITECTURE.md
index 56c28f09..9ffb8b4d 100644
--- a/ARCHITECTURE.md
+++ b/ARCHITECTURE.md
@@ -300,6 +300,13 @@ Lifecycle behavior is covered by
| [config/admin](src/free_claude_code/config/admin/) | Catalog-driven fields, presentation, validation and sparse persistence |
| [runtime/application.py](src/free_claude_code/runtime/application.py) | Credential checks, Apply publication/restart decisions and account operations |
+Retired provider references are normalized per source before validation by
+`config/model_refs.py`. Managed selections are cleared to inherit the default;
+process overrides retain their ownership. Startup atomically repairs managed
+config and current Chat model selections; Chat history retains its original model
+facts. Stale direct request IDs route to the effective default, preserving encoded
+reasoning intent. Unknown preserved config values are masked in Admin previews.
+
Live precedence is defaults, managed `~/.fcc/.env`, then process environment.
The deliberate exception is a non-empty managed `ANTHROPIC_AUTH_TOKEN`, which
wins over an inherited token. Settings loading may perform the one-time legacy
diff --git a/README.md b/README.md
index 46de2efe..fb040a92 100644
--- a/README.md
+++ b/README.md
@@ -25,7 +25,7 @@
## What You Get
-- **50 ToS-friendly providers. 1.3B+ free tokens every month.** Use free, paid, subscription, and local models from one searchable UI without putting your account at risk. FCC follows provider terms and removes integrations if they stop being allowed.
+- **49 ToS-friendly providers. 1.3B+ free tokens every month.** Use free, paid, subscription, and local models from one searchable UI without putting your account at risk. FCC follows provider terms and removes integrations if they stop being allowed.
- **10 coding agents. One model catalog.** Run [Claude Code](https://code.claude.com/docs/en/overview), [Codex](https://github.com/openai/codex), [Pi](https://github.com/earendil-works/pi), [OpenCode](https://github.com/anomalyco/opencode), [Cline](https://github.com/cline/cline), [Hermes](https://github.com/NousResearch/hermes-agent), [DeepSeek Harness](https://github.com/deepseek-ai/deepseek-harness), [Grok Build](https://github.com/xai-org/grok-build), [Muse Code](https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2/), or [Aider](https://aider.chat/) with your FCC models.
- **Keep coding through provider outages.** After retries are exhausted, FCC automatically tries your next configured model without making you restart the turn—across every client.
- **Up to 90% fewer terminal-output tokens.** Optional [RTK](https://github.com/rtk-ai/rtk) filters common command output, while five FCC optimizations handle quota probes, command-prefix detection, titles, suggestions, and filepaths without calling a provider.
@@ -185,6 +185,11 @@ from more than one provider before succeeding.
<details>
<summary><strong>Provider catalog</strong></summary>
+[GitHub Models retired on July 30, 2026](https://github.blog/changelog/2026-07-30-github-models-is-now-retired/). FCC automatically resets retired model
+selections to your configured default, or the built-in default when `MODEL` itself
+is retired. Configure that default provider in Admin if its credentials are missing.
+GitHub Copilot is a separate service and is not a replacement provider in FCC.
+
| Provider | Admin UI setting | Example `MODEL` |
| --- | --- | --- |
| [NVIDIA NIM](https://build.nvidia.com/settings/api-keys) | `NVIDIA_NIM_API_KEY` | `nvidia_nim/nvidia/nemotron-3-super-120b-a12b` |
@@ -216,7 +221,6 @@ from more than one provider before succeeding.
| [Amazon Bedrock](https://console.aws.amazon.com/bedrock/) | `AWS_BEARER_TOKEN_BEDROCK` | `bedrock/openai.gpt-oss-120b` |
| [Hugging Face Inference Providers](https://huggingface.co/settings/tokens) | `HUGGINGFACE_API_KEY` | `huggingface/Qwen/Qwen3-Coder-480B-A35B-Instruct:fastest` |
| [Cohere](https://dashboard.cohere.com/api-keys) | `COHERE_API_KEY` | `cohere/command-a-plus-05-2026` |
-| [GitHub Models](https://github.com/marketplace?type=models) | `GITHUB_MODELS_TOKEN` | `github_models/openai/gpt-4.1` |
| [Wafer](https://wafer.ai/) | `WAFER_API_KEY` | `wafer/DeepSeek-V4-Pro` |
| [Kimi API](https://platform.moonshot.ai/console/api-keys) | `KIMI_API_KEY` | `kimi/kimi-k2.5` |
| [Kimi Code](https://www.kimi.com/code/console) | `KIMI_CODE_API_KEY` | `kimi_code/k3` |
diff --git a/e2e/test_admin_providers.py b/e2e/test_admin_providers.py
index 16aff0a1..3b1705e4 100644
--- a/e2e/test_admin_providers.py
+++ b/e2e/test_admin_providers.py
@@ -4,6 +4,25 @@ import pytest
from playwright.sync_api import ConsoleMessage, Page, ViewportSize, expect
+ "admin_base_url", [{"MODEL": "github_models/openai/old"}], indirect=True
+)
+def test_retired_provider_is_absent_and_default_setup_remains_available(
+ page: Page, admin_base_url: str
+):
+ _open_admin(page, admin_base_url, {"width": 1280, "height": 720})
+ expect(page.locator('[data-provider="github_models"]')).to_have_count(0)
+ expect(page.locator("#field-GITHUB_MODELS_TOKEN")).to_have_count(0)
+ card = page.locator('[data-provider="nvidia_nim"]')
+ expect(card.locator(".status-pill")).to_have_text("Missing key")
+ card.get_by_role("button", name="Configure", exact=True).click()
+ expect(page.locator("#field-NVIDIA_NIM_API_KEY")).to_be_focused()
+ page.get_by_role("button", name="Model Config", exact=True).click()
+ expect(page.locator("#field-MODEL")).to_have_value(
+ "nvidia_nim/nvidia/nemotron-3-super-120b-a12b"
+ )
+
+
def _open_admin(
page: Page,
admin_base_url: str,
diff --git a/e2e/test_chat_sessions.py b/e2e/test_chat_sessions.py
index db610585..89c9db2b 100644
--- a/e2e/test_chat_sessions.py
+++ b/e2e/test_chat_sessions.py
@@ -35,6 +35,25 @@ def _hold_next_chat_operation(page: Page, action: str) -> None:
)
+def test_stale_chat_selection_reloads_as_default(page: Page, admin_base_url: str):
+ session = page.request.post(
+ f"{admin_base_url}/admin/api/chat/sessions", data={}
+ ).json()
+ response = page.request.patch(
+ f"{admin_base_url}/admin/api/chat/sessions/{session['id']}",
+ data={
+ "expected_revision": session["revision"],
+ "model": "github_models/openai/old",
+ },
+ )
+ assert response.ok
+ page.goto(f"{admin_base_url}/admin/chat/{session['id']}")
+ expect(page.get_by_role("combobox", name="Selected model")).to_have_value(
+ "open_router/e2e-default"
+ )
+ expect(page.get_by_role("textbox", name="Message", exact=True)).to_be_visible()
+
+
def _select_model(page: Page, model_ref: str) -> None:
model = page.get_by_role("combobox", name="Selected model")
model.click()
diff --git a/pyproject.toml b/pyproject.toml
index b4caf609..5f72706d 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
[project]
name = "free-claude-code"
-version = "5.20.0"
+version = "5.21.0"
description = "Local proxy connecting coding agents to OpenAI-compatible AI providers"
readme = "README.md"
requires-python = ">=3.14.0"
diff --git a/smoke/lib/config.py b/smoke/lib/config.py
index 1cf6ccf5..8d35f5dd 100644
--- a/smoke/lib/config.py
+++ b/smoke/lib/config.py
@@ -68,7 +68,6 @@ PROVIDER_SMOKE_DEFAULT_MODELS: dict[str, str] = {
"bedrock": "bedrock/openai.gpt-oss-120b",
"huggingface": "huggingface/openai/gpt-oss-120b:fastest",
"cohere": "cohere/command-a-plus-05-2026",
- "github_models": "github_models/openai/gpt-4.1",
"zai": "zai/glm-5.2",
"zai_api": "zai_api/glm-4.7-flash",
"gemini": "gemini/models/gemini-3.1-flash-lite",
diff --git a/src/free_claude_code/application/chat/ports.py b/src/free_claude_code/application/chat/ports.py
index b5777732..fe132aaa 100644
--- a/src/free_claude_code/application/chat/ports.py
+++ b/src/free_claude_code/application/chat/ports.py
@@ -31,6 +31,14 @@ class ChatStorePort(Protocol):
async def close(self) -> None: ...
+ async def repair_retired_model_selections(
+ self,
+ *,
+ retired_provider_ids: frozenset[str],
+ default_model: str,
+ disable_reasoning: bool,
+ ) -> int: ...
+
async def load_preferences(self) -> ChatPreferences: ...
async def save_system_prompt(self, system_prompt: str) -> ChatPreferences: ...
diff --git a/src/free_claude_code/application/chat/service.py b/src/free_claude_code/application/chat/service.py
index 82a3b04e..fe52031a 100644
--- a/src/free_claude_code/application/chat/service.py
+++ b/src/free_claude_code/application/chat/service.py
@@ -13,6 +13,10 @@ from loguru import logger
from free_claude_code.application.execution import ProviderExecutor
from free_claude_code.application.ports import RequestRuntimeLease, RequestRuntimePort
from free_claude_code.application.routing import ProviderModelTarget
+from free_claude_code.config.model_refs import (
+ RETIRED_PROVIDER_IDS,
+ is_retired_model_ref,
+)
from free_claude_code.core.anthropic import (
ContentBlockText,
MessagesResponse,
@@ -124,12 +128,61 @@ class ChatService:
async def start(self) -> None:
if self._started:
- return
+ if self._accepting:
+ return
+ # A prior startup cleanup failed. Release its store before retrying.
+ retry_cleanup = asyncio.create_task(self._close_startup_store())
+ cancellation = await _wait_task_despite_cancellation(retry_cleanup)
+ try:
+ retry_cleanup.result()
+ except Exception as exc:
+ if cancellation is not None:
+ raise cancellation from exc
+ raise
+ if cancellation is not None:
+ raise cancellation
+ cancellation: asyncio.CancelledError | None = None
try:
await self._store.start()
- except asyncio.CancelledError:
- raise
- except Exception as exc:
+ self._started = True
+ default_model = self._runtime.current_settings().model
+ option = self._context.model(default_model)
+ repair_task = asyncio.create_task(
+ self._store.repair_retired_model_selections(
+ retired_provider_ids=RETIRED_PROVIDER_IDS,
+ default_model=default_model,
+ disable_reasoning=option.supports_reasoning is False,
+ ),
+ name="fcc-chat-retirement-repair",
+ )
+ cancellation = await _wait_task_despite_cancellation(repair_task)
+ repaired = repair_task.result()
+ if cancellation is not None:
+ raise cancellation
+ if repaired:
+ logger.info(
+ "Repaired retired model selections in {} Chat sessions", repaired
+ )
+ except BaseException as exc:
+ if self._started:
+ try:
+ cleanup_task = asyncio.create_task(
+ self._close_startup_store(), name="fcc-chat-start-cleanup"
+ )
+ cleanup_cancellation = await _wait_task_despite_cancellation(
+ cleanup_task
+ )
+ cancellation = cancellation or cleanup_cancellation
+ cleanup_task.result()
+ except Exception as cleanup_exc:
+ logger.warning(
+ "Chat startup cleanup failed: exc_type={}",
+ type(cleanup_exc).__name__,
+ )
+ if not isinstance(exc, Exception):
+ raise
+ if cancellation is not None:
+ raise cancellation from exc
self._unavailable_message = (
str(exc)
if isinstance(exc, ChatUnavailableError)
@@ -144,6 +197,10 @@ class ChatService:
self._accepting = True
self._unavailable_message = None
+ async def _close_startup_store(self) -> None:
+ await self._store.close()
+ self._started = False
+
async def close(self) -> None:
self._accepting = False
async with self._active_lock:
@@ -289,7 +346,12 @@ class ChatService:
if len(title) > 200:
raise ChatValidationError("Chat title cannot exceed 200 characters.")
if model is not None:
- self._context.model(model)
+ retired = is_retired_model_ref(model)
+ if retired:
+ model = self._runtime.current_settings().model
+ option = self._context.model(model)
+ if retired and option.supports_reasoning is False:
+ reasoning = ChatReasoning.OFF
async with self._active_lock:
if session_id in self._deleting:
raise ChatConflictError("This chat is being deleted.")
diff --git a/src/free_claude_code/application/routing.py b/src/free_claude_code/application/routing.py
index 86b57f2c..7ca3ed46 100644
--- a/src/free_claude_code/application/routing.py
+++ b/src/free_claude_code/application/routing.py
@@ -5,7 +5,11 @@ from dataclasses import dataclass
from loguru import logger
from free_claude_code.application.errors import UnknownProviderError
-from free_claude_code.config.model_refs import parse_model_name, parse_provider_type
+from free_claude_code.config.model_refs import (
+ is_retired_model_ref,
+ parse_model_name,
+ parse_provider_type,
+)
from free_claude_code.config.provider_catalog import (
PROVIDER_CATALOG,
SUPPORTED_PROVIDER_IDS,
@@ -147,6 +151,17 @@ class ModelRouter:
self, model_name: str
) -> tuple[str | None, str | None, bool]:
decoded = decode_gateway_model_id(model_name)
+ candidate = (
+ f"{decoded.provider_id}/{decoded.provider_model}"
+ if decoded is not None
+ else model_name
+ )
+ if is_retired_model_ref(candidate):
+ return (
+ parse_provider_type(self._settings.model),
+ parse_model_name(self._settings.model),
+ decoded.force_reasoning_off if decoded is not None else False,
+ )
if decoded is not None:
if decoded.provider_id not in SUPPORTED_PROVIDER_IDS:
return None, None, False
diff --git a/src/free_claude_code/config/admin/persistence.py b/src/free_claude_code/config/admin/persistence.py
index 12d2d0d7..8d4d1aee 100644
--- a/src/free_claude_code/config/admin/persistence.py
+++ b/src/free_claude_code/config/admin/persistence.py
@@ -14,6 +14,7 @@ from free_claude_code.config.env_migrations import (
recognized_env_keys,
render_managed_config,
)
+from free_claude_code.config.model_refs import normalize_retired_model_settings
from free_claude_code.config.paths import managed_env_path
from free_claude_code.config.provider_proxies import invalid_provider_proxy_keys
from free_claude_code.config.settings import Settings
@@ -103,7 +104,7 @@ def target_values_with_updates(
values[key] = value
values[FCC_CONFIG_SCHEMA_ENV] = CONFIG_SCHEMA_VERSION
- return values
+ return normalize_retired_model_settings(values, preserve_empty_overrides=False)
def changed_pending_fields(
diff --git a/src/free_claude_code/config/admin/provider_manifest.py b/src/free_claude_code/config/admin/provider_manifest.py
index d0f584cb..0befa6d5 100644
--- a/src/free_claude_code/config/admin/provider_manifest.py
+++ b/src/free_claude_code/config/admin/provider_manifest.py
@@ -89,13 +89,6 @@ _PROVIDER_FIELD_OVERRIDES: dict[str, ProviderFieldOverride] = {
"label": "Cohere API Key",
"description": "Cohere API key for the OpenAI-compatible Compatibility API.",
},
- "GITHUB_MODELS_TOKEN": {
- "label": "GitHub Models Token",
- "description": (
- "GitHub token with Models access for the OpenAI-compatible inference API "
- "at models.github.ai."
- ),
- },
"ZAI_API_KEY": {
"label": "Z.ai API Key",
"description": (
diff --git a/src/free_claude_code/config/constants.py b/src/free_claude_code/config/constants.py
index c4ea854a..d96a1617 100644
--- a/src/free_claude_code/config/constants.py
+++ b/src/free_claude_code/config/constants.py
@@ -1,5 +1,7 @@
"""Shared defaults used by config models and provider adapters."""
+DEFAULT_MODEL = "nvidia_nim/nvidia/nemotron-3-super-120b-a12b"
+
# HTTP client connect timeout (seconds). Keep aligned with README.md and .env.example.
HTTP_CONNECT_TIMEOUT_DEFAULT = 10.0
diff --git a/src/free_claude_code/config/env_migrations.py b/src/free_claude_code/config/env_migrations.py
index d9c84f63..896094cd 100644
--- a/src/free_claude_code/config/env_migrations.py
+++ b/src/free_claude_code/config/env_migrations.py
@@ -6,6 +6,8 @@ from collections.abc import Mapping
from dataclasses import dataclass
from pathlib import Path
+from loguru import logger
+
from .env_files import (
ANTHROPIC_AUTH_TOKEN_ENV,
FCC_CONFIG_SCHEMA_ENV,
@@ -15,6 +17,7 @@ from .env_files import (
legacy_explicit_env_path,
verified_checkout_env_path,
)
+from .model_refs import normalize_retired_model_settings
from .paths import legacy_env_paths, managed_env_path
from .provider_catalog import PROVIDER_CATALOG
from .settings import Settings
@@ -190,7 +193,14 @@ def consolidate_managed_config(
if base_is_managed:
schema = base_values.get(FCC_CONFIG_SCHEMA_ENV)
if schema == CONFIG_SCHEMA_VERSION:
- return ConfigMigrationResult(changed=False)
+ normalized = normalize_retired_model_settings(
+ base_values, preserve_empty_overrides=False
+ )
+ if normalized == base_values:
+ return ConfigMigrationResult(changed=False)
+ atomic_write_managed_config(normalized, path=managed)
+ _log_model_repairs(base_values, normalized)
+ return ConfigMigrationResult(changed=True)
if schema not in {None, ""}:
raise ValueError(
f"Managed config {managed} uses unsupported schema {schema!r}; "
@@ -224,11 +234,23 @@ def consolidate_managed_config(
if key in known and key != FCC_CONFIG_SCHEMA_ENV and not values[key].strip():
values.pop(key)
values[FCC_CONFIG_SCHEMA_ENV] = CONFIG_SCHEMA_VERSION
-
- atomic_write_managed_config(values, path=managed)
+ normalized = normalize_retired_model_settings(
+ values, preserve_empty_overrides=False
+ )
+ atomic_write_managed_config(normalized, path=managed)
+ _log_model_repairs(values, normalized)
return ConfigMigrationResult(changed=True, imported_from=tuple(imported))
+def _log_model_repairs(before: Mapping[str, str], after: Mapping[str, str]) -> None:
+ changed = sorted(key for key in before if before[key] != after.get(key))
+ if changed:
+ logger.info(
+ "Repaired retired provider selections in managed config: {}",
+ ", ".join(changed),
+ )
+
+
def atomic_write_managed_config(
values: Mapping[str, str],
*,
@@ -292,7 +314,8 @@ def render_managed_config(
if unknown:
lines.extend(("", "# Preserved unrecognized settings"))
lines.extend(
- f"{key}={quote_env_value(canonical[key])}" for key in sorted(unknown)
+ f"{key}={quote_env_value('********' if mask_secrets else canonical[key])}"
+ for key in sorted(unknown)
)
return "\n".join(lines) + "\n"
diff --git a/src/free_claude_code/config/loader.py b/src/free_claude_code/config/loader.py
index 05d0e683..569bee6f 100644
--- a/src/free_claude_code/config/loader.py
+++ b/src/free_claude_code/config/loader.py
@@ -15,6 +15,7 @@ from .env_migrations import (
consolidate_managed_config,
settings_env_keys,
)
+from .model_refs import normalize_retired_model_settings
from .paths import config_lock_path, managed_env_path
from .provider_proxies import invalid_provider_proxy_keys
from .settings import Settings
@@ -97,7 +98,10 @@ def compose_settings_snapshot(
) -> SettingsSnapshot:
"""Validate prospective managed values with live process precedence."""
- process = env if env is not None else os.environ
+ process = normalize_retired_model_settings(
+ env if env is not None else os.environ, preserve_empty_overrides=True
+ )
+ managed = normalize_retired_model_settings(managed, preserve_empty_overrides=False)
aliases = _settings_aliases()
recognized = settings_env_keys()
values: dict[str, str] = {
diff --git a/src/free_claude_code/config/model_refs.py b/src/free_claude_code/config/model_refs.py
index fb4487b0..2491caa2 100644
--- a/src/free_claude_code/config/model_refs.py
+++ b/src/free_claude_code/config/model_refs.py
@@ -1,8 +1,55 @@
"""Provider-prefixed model reference helpers."""
+from collections.abc import Mapping
from dataclasses import dataclass
from typing import Protocol
+from .constants import DEFAULT_MODEL
+
+RETIRED_PROVIDER_IDS = frozenset({"github_models"})
+
+
+def is_retired_model_ref(model_ref: str) -> bool:
+ """Recognize complete references owned by a retired provider."""
+
+ provider, separator, model = model_ref.strip().partition("/")
+ return provider in RETIRED_PROVIDER_IDS and bool(separator and model.strip())
+
+
+def parse_model_fallbacks(value: object) -> object:
+ if value is None:
+ return None
+ if isinstance(value, str):
+ if not value.strip():
+ return None
+ return tuple(part.strip() for part in value.split(","))
+ if isinstance(value, list | tuple) and not value:
+ return None
+ return value
+
+
+def normalize_retired_model_settings(
+ values: Mapping[str, str], *, preserve_empty_overrides: bool
+) -> dict[str, str]:
+ """Repair one source without changing its precedence or unrelated validation."""
+
+ normalized = dict(values)
+ for key in ("MODEL", "MODEL_FABLE", "MODEL_OPUS", "MODEL_SONNET", "MODEL_HAIKU"):
+ if is_retired_model_ref(normalized.get(key, "")):
+ if preserve_empty_overrides:
+ normalized[key] = DEFAULT_MODEL if key == "MODEL" else ""
+ else:
+ normalized.pop(key)
+ fallbacks = parse_model_fallbacks(normalized.get("MODEL_FALLBACKS"))
+ if isinstance(fallbacks, tuple) and all(fallbacks):
+ retained = tuple(ref for ref in fallbacks if not is_retired_model_ref(ref))
+ if retained != fallbacks:
+ if retained or preserve_empty_overrides:
+ normalized["MODEL_FALLBACKS"] = ",".join(retained)
+ else:
+ normalized.pop("MODEL_FALLBACKS")
+ return normalized
+
@dataclass(frozen=True, slots=True)
class ConfiguredChatModelRef:
diff --git a/src/free_claude_code/config/provider_catalog.py b/src/free_claude_code/config/provider_catalog.py
index 8fb63945..a63d8bf3 100644
--- a/src/free_claude_code/config/provider_catalog.py
+++ b/src/free_claude_code/config/provider_catalog.py
@@ -37,7 +37,6 @@ VERCEL_AI_GATEWAY_DEFAULT_BASE = "https://ai-gateway.vercel.sh/v1"
BEDROCK_DEFAULT_BASE = "https://bedrock-mantle.us-east-1.api.aws/v1"
HUGGINGFACE_DEFAULT_BASE = "https://router.huggingface.co/v1"
COHERE_DEFAULT_BASE = "https://api.cohere.ai/compatibility/v1"
-GITHUB_MODELS_DEFAULT_BASE = "https://models.github.ai/inference"
# Z.ai OpenAI-compatible Chat Completions APIs. The endpoint selects billing.
ZAI_CODING_DEFAULT_BASE = "https://api.z.ai/api/coding/paas/v4"
ZAI_API_DEFAULT_BASE = "https://api.z.ai/api/paas/v4"
@@ -390,15 +389,6 @@ PROVIDER_CATALOG: dict[str, ProviderDescriptor] = {
default_base_url=COHERE_DEFAULT_BASE,
proxy_attr="cohere_proxy",
),
- "github_models": ProviderDescriptor(
- provider_id="github_models",
- display_name="GitHub Models",
- credential_env="GITHUB_MODELS_TOKEN",
- credential_url="https://github.com/settings/tokens",
- credential_attr="github_models_token",
- default_base_url=GITHUB_MODELS_DEFAULT_BASE,
- proxy_attr="github_models_proxy",
- ),
"wafer": ProviderDescriptor(
provider_id="wafer",
display_name="Wafer",
@@ -589,7 +579,7 @@ PROVIDER_CATALOG: dict[str, ProviderDescriptor] = {
# Key order:
# NVIDIA NIM, OpenRouter, and Groq lead the customer-facing ranking;
# OpenCode gateways remain adjacent,
-# Vercel / Hugging Face / Cohere / GitHub Models follow gateway-style remotes,
+# Vercel / Hugging Face / Cohere follow gateway-style remotes,
# then cloud gateways, Ollama Cloud, and local providers per project plan
# (github.com/cheahjs/free-llm-api-resources Free Providers TOC as rough guide
# beyond fixed slots).
diff --git a/src/free_claude_code/config/settings.py b/src/free_claude_code/config/settings.py
index ff4b7d14..18e5f5e5 100644
--- a/src/free_claude_code/config/settings.py
+++ b/src/free_claude_code/config/settings.py
@@ -12,7 +12,8 @@ from pydantic import (
model_validator,
)
-from .constants import HTTP_CONNECT_TIMEOUT_DEFAULT
+from .constants import DEFAULT_MODEL, HTTP_CONNECT_TIMEOUT_DEFAULT
+from .model_refs import parse_model_fallbacks
from .nim import NimSettings
from .provider_catalog import (
BEDROCK_DEFAULT_BASE,
@@ -29,18 +30,6 @@ def _empty_to_none(value: object) -> object:
return value
-def _parse_model_fallbacks(value: object) -> object:
- if value is None:
- return None
- if isinstance(value, str):
- if not value.strip():
- return None
- return tuple(part.strip() for part in value.split(","))
- if isinstance(value, list | tuple) and not value:
- return None
- return value
-
-
NonEmptyString = Annotated[
str,
StringConstraints(strip_whitespace=True, min_length=1),
@@ -51,7 +40,7 @@ OptionalNonEmptyString = Annotated[
]
OptionalModelFallbacks = Annotated[
tuple[NonEmptyString, ...] | None,
- BeforeValidator(_parse_model_fallbacks),
+ BeforeValidator(parse_model_fallbacks),
]
@@ -158,11 +147,6 @@ class Settings(BaseModel):
default=None, validation_alias="COHERE_API_KEY"
)
- # ==================== GitHub Models ====================
- github_models_token: OptionalNonEmptyString = Field(
- default=None, validation_alias="GITHUB_MODELS_TOKEN"
- )
-
# ==================== SambaNova Cloud ====================
sambanova_api_key: OptionalNonEmptyString = Field(
default=None, validation_alias="SAMBANOVA_API_KEY"
@@ -357,7 +341,7 @@ class Settings(BaseModel):
# All Claude model requests are mapped to this single model (fallback)
# Format: provider_type/model/name
model: NonEmptyString = Field(
- default="nvidia_nim/nvidia/nemotron-3-super-120b-a12b",
+ default=DEFAULT_MODEL,
validation_alias="MODEL",
)
@@ -471,9 +455,6 @@ class Settings(BaseModel):
cohere_proxy: OptionalNonEmptyString = Field(
default=None, validation_alias="COHERE_PROXY"
)
- github_models_proxy: OptionalNonEmptyString = Field(
- default=None, validation_alias="GITHUB_MODELS_PROXY"
- )
sambanova_proxy: OptionalNonEmptyString = Field(
default=None, validation_alias="SAMBANOVA_PROXY"
)
diff --git a/src/free_claude_code/providers/credential_validation.py b/src/free_claude_code/providers/credential_validation.py
index 5dc892c0..3cf64eab 100644
--- a/src/free_claude_code/providers/credential_validation.py
+++ b/src/free_claude_code/providers/credential_validation.py
@@ -155,7 +155,6 @@ _PROBES = (
_Probe("deepinfra", "https://api.deepinfra.com/v1/me", _identity("uid")),
_Probe("mistral", "/models", _MODELS),
_Probe("wandb", "/models", _MODELS),
- _Probe("github_models", "https://api.github.com/user", _identity("login")),
_Probe("minimax", "/models", _MODELS),
_Probe(
"novita",
@@ -260,8 +259,6 @@ async def _check_one(settings: Settings, key: str, probe: _Probe) -> CredentialC
headers["x-goog-api-key"] = credential
else:
headers["Authorization"] = f"Bearer {credential}"
- if probe.provider_id == "github_models":
- headers["Accept"] = "application/vnd.github+json"
if probe.provider_id == "novita":
headers["Content-Type"] = "application/json"
try:
diff --git a/src/free_claude_code/providers/github_models/__init__.py b/src/free_claude_code/providers/github_models/__init__.py
deleted file mode 100644
index d86c4d6f..00000000
--- a/src/free_claude_code/providers/github_models/__init__.py
+++ /dev/null
@@ -1,5 +0,0 @@
-"""GitHub Models provider."""
-
-from .client import GitHubModelsProvider
-
-__all__ = ["GitHubModelsProvider"]
diff --git a/src/free_claude_code/providers/github_models/client.py b/src/free_claude_code/providers/github_models/client.py
deleted file mode 100644
index 639d3fd9..00000000
--- a/src/free_claude_code/providers/github_models/client.py
+++ /dev/null
@@ -1,155 +0,0 @@
-"""GitHub Models provider using OpenAI-compatible chat completions."""
-
-from collections.abc import Mapping, Sequence
-from typing import Any
-
-import httpx
-
-from free_claude_code.application.model_metadata import ProviderModelInfo
-from free_claude_code.core.anthropic import ReasoningReplayMode
-from free_claude_code.providers.admission import (
- ProviderAdmissionController,
- ProviderOperationKind,
-)
-from free_claude_code.providers.base import ProviderConfig
-from free_claude_code.providers.http import maybe_await_aclose
-from free_claude_code.providers.model_listing import (
- ModelListResponseError,
- model_infos_from_ids,
-)
-from free_claude_code.providers.openai_chat import (
- NO_REASONING,
- OpenAIChatProfile,
- OpenAIChatProvider,
- OpenAIChatRequestPolicy,
-)
-
-GITHUB_MODELS_CATALOG_URL = "https://models.github.ai/catalog/models"
-GITHUB_MODELS_API_VERSION = "2026-03-10"
-
-_REQUEST_POLICY = OpenAIChatRequestPolicy(
- provider_name="GITHUB_MODELS",
- reasoning_replay=ReasoningReplayMode.THINK_TAGS,
-)
-_PROFILE = OpenAIChatProfile(_REQUEST_POLICY, NO_REASONING)
-_REQUIRED_MODEL_CAPABILITIES = frozenset({"streaming", "tool-calling"})
-
-
-class GitHubModelsProvider(OpenAIChatProvider):
- """GitHub Models OpenAI-compatible inference provider."""
-
- def __init__(
- self, config: ProviderConfig, *, admission: ProviderAdmissionController
- ):
- self._catalog_url = GITHUB_MODELS_CATALOG_URL
- self._model_list_client = httpx.AsyncClient(
- proxy=config.proxy,
- timeout=httpx.Timeout(
- config.http_read_timeout,
- connect=config.http_connect_timeout,
- read=config.http_read_timeout,
- write=config.http_write_timeout,
- ),
- )
- super().__init__(
- config,
- profile=_PROFILE,
- admission=admission,
- default_headers=_github_models_default_headers(),
- )
-
- async def cleanup(self) -> None:
- """Release provider client resources."""
- await super().cleanup()
- await self._model_list_client.aclose()
-
- async def list_model_infos(self) -> frozenset[ProviderModelInfo]:
- """Return stream/tool-capable GitHub Models catalog ids."""
-
- async def request() -> httpx.Response:
- response = await self._model_list_client.get(
- self._catalog_url,
- headers=self._model_list_headers(),
- )
- try:
- response.raise_for_status()
- except Exception:
- await maybe_await_aclose(response)
- raise
- return response
-
- execution = self._admission.start_execution()
- response = await execution.run_call(
- request,
- operation_kind=ProviderOperationKind.MODEL_DISCOVERY,
- )
- try:
- try:
- payload = response.json()
- except ValueError as exc:
- raise ModelListResponseError(
- "GITHUB_MODELS model-list response is malformed: invalid JSON"
- ) from exc
- return model_infos_from_ids(
- _extract_supported_github_model_ids(payload),
- )
- finally:
- await maybe_await_aclose(response)
-
- def _model_list_headers(self) -> dict[str, str]:
- if self._api_key is None:
- raise AssertionError("GitHub Models requires a static API key")
- return _github_models_api_headers(self._api_key)
-
-
-def _github_models_default_headers() -> dict[str, str]:
- return {
- "Accept": "application/vnd.github+json",
- "X-GitHub-Api-Version": GITHUB_MODELS_API_VERSION,
- }
-
-
-def _github_models_api_headers(api_key: str) -> dict[str, str]:
- return {
- **_github_models_default_headers(),
- "Authorization": f"Bearer {api_key}",
- }
-
-
-def _extract_supported_github_model_ids(payload: Any) -> frozenset[str]:
- """Extract stream/tool-capable model ids from GitHub's catalog array."""
- if not _is_sequence(payload):
- raise ModelListResponseError(
- "GITHUB_MODELS model-list response is malformed: expected top-level array"
- )
-
- model_ids: set[str] = set()
- for item in payload:
- if not isinstance(item, Mapping):
- raise ModelListResponseError(
- "GITHUB_MODELS model-list response is malformed: expected every item to be an object"
- )
- model_id = item.get("id")
- if not isinstance(model_id, str) or not model_id.strip():
- raise ModelListResponseError(
- "GITHUB_MODELS model-list response is malformed: expected every item to include id"
- )
- capabilities = item.get("capabilities")
- if not _supports_streaming_tools(capabilities):
- continue
- model_ids.add(model_id)
-
- return frozenset(model_ids)
-
-
-def _supports_streaming_tools(capabilities: Any) -> bool:
- if not _is_sequence(capabilities):
- return False
- capability_names = {item for item in capabilities if isinstance(item, str)}
- return capability_names >= _REQUIRED_MODEL_CAPABILITIES
-
-
-def _is_sequence(value: Any) -> bool:
- return isinstance(value, Sequence) and not isinstance(
- value, str | bytes | bytearray
- )
diff --git a/src/free_claude_code/providers/runtime/factory.py b/src/free_claude_code/providers/runtime/factory.py
index 1434ae3b..079854f1 100644
--- a/src/free_claude_code/providers/runtime/factory.py
+++ b/src/free_claude_code/providers/runtime/factory.py
@@ -125,16 +125,6 @@ def _create_vertex(
)
-def _create_github_models(
- config: ProviderConfig,
- _settings: Settings,
- admission: ProviderAdmissionController,
-) -> BaseProvider:
- from free_claude_code.providers.github_models import GitHubModelsProvider
-
- return GitHubModelsProvider(config, admission=admission)
-
-
def _create_groq(
config: ProviderConfig,
_settings: Settings,
@@ -175,7 +165,6 @@ _SPECIAL_PROVIDER_FACTORIES: dict[str, ProviderFactory] = {
"cloudflare": _create_cloudflare,
"gemini": _create_gemini,
"vertex": _create_vertex,
- "github_models": _create_github_models,
"groq": _create_groq,
"opencode_zen": _create_opencode_zen,
"opencode_go": _create_opencode_go,
diff --git a/src/free_claude_code/runtime/chat_sqlite.py b/src/free_claude_code/runtime/chat_sqlite.py
index 630da8fd..43174e4e 100644
--- a/src/free_claude_code/runtime/chat_sqlite.py
+++ b/src/free_claude_code/runtime/chat_sqlite.py
@@ -89,6 +89,61 @@ class SQLiteChatStore:
async def load_preferences(self) -> ChatPreferences:
return await self._run(self._load_preferences)
+ async def repair_retired_model_selections(
+ self,
+ *,
+ retired_provider_ids: frozenset[str],
+ default_model: str,
+ disable_reasoning: bool,
+ ) -> int:
+ """Repair current selections atomically, preserving historical model facts."""
+
+ def operation(connection: sqlite3.Connection) -> int:
+ connection.execute("BEGIN IMMEDIATE")
+ repaired = 0
+ now = _now_ms()
+ for provider_id in sorted(retired_provider_ids):
+ prefix = f"{provider_id}/"
+ cursor = connection.execute(
+ """
+ UPDATE chat_sessions
+ SET model = ?, reasoning = CASE WHEN ? THEN 'off' ELSE reasoning END,
+ revision = revision + 1, updated_at = ?
+ WHERE substr(model, 1, ?) = ? AND length(trim(substr(model, ?))) > 0
+ """,
+ (
+ default_model,
+ disable_reasoning,
+ now,
+ len(prefix),
+ prefix,
+ len(prefix) + 1,
+ ),
+ )
+ repaired += cursor.rowcount
+ connection.execute(
+ """
+ UPDATE chat_settings
+ SET last_model = ?,
+ last_reasoning = CASE WHEN ? THEN 'off' ELSE last_reasoning END,
+ updated_at = ?
+ WHERE substr(last_model, 1, ?) = ?
+ AND length(trim(substr(last_model, ?))) > 0
+ """,
+ (
+ default_model,
+ disable_reasoning,
+ now,
+ len(prefix),
+ prefix,
+ len(prefix) + 1,
+ ),
+ )
+ connection.commit()
+ return repaired
+
+ return await self._run(operation)
+
async def save_system_prompt(self, system_prompt: str) -> ChatPreferences:
def operation(connection: sqlite3.Connection) -> ChatPreferences:
now = _now_ms()
diff --git a/tests/api/test_admin.py b/tests/api/test_admin.py
index 3a2629ef..047cca6d 100644
--- a/tests/api/test_admin.py
+++ b/tests/api/test_admin.py
@@ -31,6 +31,45 @@ def _local_client(app):
)
+def test_admin_retirement_preview_apply_and_runtime_agree(monkeypatch, tmp_path):
+ _set_home(monkeypatch, tmp_path)
+ _clear_process_config(monkeypatch)
+ env_file = tmp_path / ".fcc" / ".env"
+ env_file.parent.mkdir(parents=True)
+ env_file.write_text(
+ "FCC_CONFIG_SCHEMA=1\nMODEL=groq/default\nMODEL_OPUS=groq/managed\n"
+ "GITHUB_MODELS_TOKEN=preserve-secret\nGITHUB_MODELS_PROXY=http://user:[email protected]\nPRIVATE=hidden\n",
+ encoding="utf-8",
+ )
+ monkeypatch.setenv("MODEL_OPUS", "github_models/retired-process")
+ app = create_test_app()
+ client = _local_client(app)
+ body = client.post(
+ "/admin/api/config/apply",
+ json={
+ "values": {
+ "MODEL_SONNET": "github_models/stale",
+ "MODEL_OPUS": "deepseek/cannot-unlock",
+ "MODEL_FALLBACKS": "groq/a,github_models/old,deepseek/b",
+ }
+ },
+ ).json()
+ assert body["applied"]
+ text = env_file.read_text(encoding="utf-8")
+ assert "MODEL=groq/default" in text
+ assert "MODEL_SONNET" not in text
+ assert "MODEL_OPUS=groq/managed" in text
+ assert "MODEL_FALLBACKS=groq/a,deepseek/b" in text
+ assert "preserve-secret" in text and "http://user:[email protected]" in text
+ assert "preserve-secret" not in body["env_preview"]
+ assert "http://user:[email protected]" not in body["env_preview"]
+ assert "PRIVATE=********" in body["env_preview"]
+ effective = provider_manager_for_app(app).current_settings()
+ assert effective.model == "groq/default"
+ assert effective.model_opus is None and effective.model_sonnet is None
+ assert effective.model_fallbacks == ("groq/a", "deepseek/b")
+
+
@pytest.fixture(autouse=True)
def _offline_credential_checks(monkeypatch):
"""These tests exercise config persistence; probe HTTP has its own suite."""
@@ -523,7 +562,7 @@ def test_admin_config_masks_secrets_and_exposes_manifest(monkeypatch, tmp_path):
assert "FIREWORKS_API_KEY" in keys
assert "CLOUDFLARE_API_TOKEN" in keys
assert "CLOUDFLARE_ACCOUNT_ID" in keys
- assert "GITHUB_MODELS_TOKEN" in keys
+ assert "GITHUB_MODELS_TOKEN" not in keys
assert "GEMINI_API_KEY" in keys
assert "GROQ_API_KEY" in keys
assert "SAMBANOVA_API_KEY" in keys
@@ -1496,7 +1535,7 @@ def test_admin_apply_writes_cohere_key_and_masks_preview(monkeypatch, tmp_path):
assert "COHERE_API_KEY=cohere-secret" in text
-def test_admin_apply_writes_github_models_token_and_masks_preview(
+def test_admin_apply_replaces_retired_model_and_ignores_retired_credential(
monkeypatch, tmp_path
):
_set_home(monkeypatch, tmp_path)
@@ -1516,11 +1555,11 @@ def test_admin_apply_writes_github_models_token_and_masks_preview(
assert response.status_code == 200
body = response.json()
assert body["applied"] is True
- assert "GITHUB_MODELS_TOKEN=********" in body["env_preview"]
+ assert "GITHUB_MODELS_TOKEN" not in body["env_preview"]
env_file = tmp_path / ".fcc" / ".env"
text = env_file.read_text(encoding="utf-8")
- assert "MODEL=github_models/openai/gpt-4.1" in text
- assert "GITHUB_MODELS_TOKEN=github-secret" in text
+ assert "MODEL=" not in text
+ assert "GITHUB_MODELS_TOKEN" not in text
def test_admin_apply_preserves_hidden_diagnostics_and_smoke_values(
diff --git a/tests/api/test_api.py b/tests/api/test_api.py
index eeb295aa..8dce6fa6 100644
--- a/tests/api/test_api.py
+++ b/tests/api/test_api.py
@@ -3,6 +3,7 @@ from unittest.mock import MagicMock, patch
import pytest
from fastapi.testclient import TestClient
+from free_claude_code.config.constants import DEFAULT_MODEL
from free_claude_code.core.anthropic import ReasoningReplayMode
from free_claude_code.core.failures import ExecutionFailure, FailureKind
from free_claude_code.core.reasoning import ReasoningPolicy
@@ -16,6 +17,30 @@ from free_claude_code.providers.openai_chat import (
from tests.api.support import create_test_app
from tests.providers.support import immediate_admission, make_provider_config
+
[email protected]("stream", [False, True])
[email protected]("prefix", ["", "anthropic/", "claude-3-freecc-no-thinking/"])
+def test_retired_model_ingress_calls_default_provider(client, stream, prefix):
+ original = f"{prefix}github_models/vendor/opus"
+ _stream_messages_calls.clear()
+ response = client.post(
+ "/v1/messages",
+ json={
+ "model": original,
+ "max_tokens": 32,
+ "messages": [{"role": "user", "content": "hello"}],
+ "stream": stream,
+ },
+ )
+ assert response.status_code == 200
+ assert len(_stream_messages_calls) == 1
+ args, kwargs = _stream_messages_calls[0]
+ assert args[0].model == DEFAULT_MODEL.split("/", 1)[1]
+ assert kwargs["response_model"] == original
+ if prefix == "claude-3-freecc-no-thinking/":
+ assert kwargs["reasoning"] == ReasoningPolicy.off()
+
+
app = create_test_app()
# Mock provider
diff --git a/tests/api/test_openai_responses.py b/tests/api/test_openai_responses.py
index 717b682d..e1811824 100644
--- a/tests/api/test_openai_responses.py
+++ b/tests/api/test_openai_responses.py
@@ -6,6 +6,7 @@ import pytest
from fastapi.testclient import TestClient
from free_claude_code.application.errors import InvalidRequestError
+from free_claude_code.config.constants import DEFAULT_MODEL
from free_claude_code.core.anthropic import ReasoningReplayMode
from free_claude_code.core.anthropic.stream_contracts import parse_sse_text
from free_claude_code.core.failures import ExecutionFailure, FailureKind
@@ -88,6 +89,19 @@ def test_responses_probe_endpoints_return_204(
assert client.options("/v1/responses").status_code == 204
[email protected]("prefix", ["", "anthropic/", "claude-3-freecc-no-thinking/"])
+def test_retired_response_id_uses_default_provider(responses_client, prefix):
+ client, provider = responses_client
+ original = f"{prefix}github_models/vendor/opus"
+ response = client.post("/v1/responses", json={"model": original, "input": "hello"})
+ assert response.status_code == 200
+ assert "response.completed" in response.text
+ assert provider.requests[0].model == DEFAULT_MODEL.split("/", 1)[1]
+ assert provider.stream_kwargs[0]["response_model"] == original
+ if prefix == "claude-3-freecc-no-thinking/":
+ assert provider.stream_kwargs[0]["reasoning"] == ReasoningPolicy.off()
+
+
def test_create_response_stream_routes_native_request_through_provider(
responses_client: tuple[TestClient, FakeProvider],
) -> None:
diff --git a/tests/api/test_routes_optimizations.py b/tests/api/test_routes_optimizations.py
index 84ec12e3..b77127fb 100644
--- a/tests/api/test_routes_optimizations.py
+++ b/tests/api/test_routes_optimizations.py
@@ -135,6 +135,33 @@ def test_count_tokens_endpoint(client):
assert response.json()["input_tokens"] == 5
[email protected]("prefix", ["", "anthropic/", "claude-3-freecc-no-thinking/"])
+def test_count_tokens_retired_model_uses_configured_default(prefix):
+ test_app = create_test_app(
+ Settings(MODEL="groq/default", MODEL_OPUS="deepseek/opus")
+ )
+ with (
+ patch("free_claude_code.api.routes.get_token_count", return_value=5),
+ patch("free_claude_code.api.handlers.token_count.trace_event") as trace,
+ ):
+ response = TestClient(test_app).post(
+ "/v1/messages/count_tokens",
+ json={
+ "model": f"{prefix}github_models/vendor/opus",
+ "messages": [{"role": "user", "content": "hi"}],
+ },
+ )
+ assert response.status_code == 200
+ assert response.json()["input_tokens"] == 5
+ routed = next(
+ call.kwargs
+ for call in trace.call_args_list
+ if call.kwargs["stage"] == "routing"
+ )
+ assert routed["provider_model_ref"] == "groq/default"
+ assert routed["gateway_model"] == f"{prefix}github_models/vendor/opus"
+
+
def test_count_tokens_error_returns_500(client):
"""When get_token_count raises, count_tokens returns 500."""
payload = {
diff --git a/tests/application/test_chat_service.py b/tests/application/test_chat_service.py
index 0235d2b3..b37a3ce6 100644
--- a/tests/application/test_chat_service.py
+++ b/tests/application/test_chat_service.py
@@ -1,5 +1,6 @@
import asyncio
import sqlite3
+import threading
from collections.abc import AsyncIterator, Awaitable, Callable
from dataclasses import dataclass
from pathlib import Path
@@ -12,6 +13,7 @@ from free_claude_code.application.chat import (
ChatNotFoundError,
ChatOperationAcknowledgement,
ChatOperationKind,
+ ChatReasoning,
ChatService,
ChatSession,
ChatUnavailableError,
@@ -23,11 +25,175 @@ from free_claude_code.config.settings import Settings
from free_claude_code.core.anthropic import MessagesRequest
from free_claude_code.core.anthropic.streaming import format_sse_event
from free_claude_code.core.failures import ExecutionFailure, FailureKind
+from free_claude_code.core.interprocess_lock import InterprocessFileLock
from free_claude_code.core.openai_responses import OpenAIResponsesRequest
from free_claude_code.core.reasoning import ReasoningPolicy
from free_claude_code.runtime.chat_sqlite import SQLiteChatStore
[email protected]("repair_fails", [False, True])
+async def test_cancelled_startup_drains_retirement_worker_before_releasing_lock(
+ tmp_path, monkeypatch, repair_fails
+):
+ store = SQLiteChatStore(tmp_path / "chat.db", tmp_path / "chat.lock")
+ await store.start()
+ original_run = store._run_sync
+ entered = threading.Event()
+ release = threading.Event()
+ finished = threading.Event()
+
+ def blocked_run(operation):
+ entered.set()
+ assert release.wait(5)
+ try:
+ if repair_fails:
+ raise OSError("injected repair failure after cancellation")
+ return original_run(operation)
+ finally:
+ finished.set()
+
+ monkeypatch.setattr(store, "_run_sync", blocked_run)
+ service = ChatService(FakeRuntime(FakeChatProvider()), store)
+ task = asyncio.create_task(service.start())
+ replacement_lock = InterprocessFileLock(tmp_path / "chat.lock")
+ try:
+ assert await asyncio.to_thread(entered.wait, 2)
+ task.cancel()
+ await asyncio.sleep(0.05)
+ assert not task.done()
+ assert not finished.is_set()
+ assert not replacement_lock.acquire()
+ release.set()
+ with pytest.raises(asyncio.CancelledError):
+ await task
+ assert finished.is_set()
+ assert replacement_lock.acquire()
+ finally:
+ release.set()
+ await asyncio.gather(task, return_exceptions=True)
+ replacement_lock.release()
+ await service.close()
+
+
[email protected]("supports_thinking", [True, False, None])
+async def test_retired_chat_selection_startup_and_stale_patch(
+ tmp_path, monkeypatch, supports_thinking
+):
+ runtime = FakeRuntime(FakeChatProvider())
+ monkeypatch.setattr(
+ runtime,
+ "cached_model_info",
+ lambda provider_id, model_id: ProviderModelInfo(
+ model_id, supports_thinking=supports_thinking
+ ),
+ )
+ store = SQLiteChatStore(tmp_path / "chat.db", tmp_path / "chat.lock")
+ await store.start()
+ old = await store.create_session(
+ session_id="aaaaaaaa-aaaa-4aaa-aaaa-aaaaaaaaaaaa",
+ model="github_models/old",
+ reasoning=ChatReasoning.HIGH,
+ )
+ await store.close()
+ service = ChatService(runtime, store)
+ await service.start()
+ try:
+ assert service.availability() == (True, None)
+ repaired = await service.get_session(old.id)
+ assert repaired.model == runtime.settings.model
+ assert repaired.revision == old.revision + 1
+ expected_reasoning = (
+ ChatReasoning.OFF if supports_thinking is False else ChatReasoning.HIGH
+ )
+ assert repaired.reasoning is expected_reasoning
+ with pytest.raises(ChatConflictError):
+ await service.update_session(
+ old.id,
+ expected_revision=old.revision,
+ title=None,
+ model="github_models/old",
+ reasoning=None,
+ )
+ assert await service.get_session(old.id) == repaired
+ # Simulate a stale tab submitting an old selection while thinking is enabled.
+ enabled = await store.update_session(
+ old.id,
+ expected_revision=repaired.revision,
+ title=None,
+ model=None,
+ reasoning=ChatReasoning.HIGH,
+ )
+ patched = await service.update_session(
+ old.id,
+ expected_revision=enabled.revision,
+ title=None,
+ model="github_models/old",
+ reasoning=None,
+ )
+ assert patched.model == runtime.settings.model
+ assert patched.reasoning is expected_reasoning
+ assert runtime.leases == []
+ finally:
+ await service.close()
+
+
[email protected]("cancel", [True, False])
+async def test_chat_retirement_failure_releases_store_lock(
+ tmp_path, monkeypatch, cancel
+):
+ store = SQLiteChatStore(tmp_path / "chat.db", tmp_path / "chat.lock")
+ service = ChatService(FakeRuntime(FakeChatProvider()), store)
+
+ async def fail_repair(**kwargs):
+ if cancel:
+ raise asyncio.CancelledError()
+ raise OSError("injected repair failure")
+
+ monkeypatch.setattr(store, "repair_retired_model_selections", fail_repair)
+ if cancel:
+ with pytest.raises(asyncio.CancelledError):
+ await service.start()
+ else:
+ await service.start()
+ assert not service.availability()[0]
+ replacement = SQLiteChatStore(tmp_path / "chat.db", tmp_path / "chat.lock")
+ await replacement.start()
+ await replacement.close()
+ await service.close()
+
+
+async def test_chat_retirement_failed_cleanup_is_retried_before_start(
+ tmp_path, monkeypatch
+):
+ store = SQLiteChatStore(tmp_path / "chat.db", tmp_path / "chat.lock")
+ service = ChatService(FakeRuntime(FakeChatProvider()), store)
+ original_repair = store.repair_retired_model_selections
+ original_close = store.close
+
+ async def fail_repair(**kwargs):
+ raise OSError("injected repair failure")
+
+ async def fail_close():
+ raise OSError("injected cleanup failure")
+
+ monkeypatch.setattr(store, "repair_retired_model_selections", fail_repair)
+ monkeypatch.setattr(store, "close", fail_close)
+ await service.start()
+ assert not service.availability()[0]
+ monkeypatch.setattr(store, "repair_retired_model_selections", original_repair)
+ monkeypatch.setattr(store, "close", original_close)
+ await service.start()
+ try:
+ assert service.availability() == (True, None)
+ await service.create_session()
+ finally:
+ await service.close()
+
+
class FakeChatProvider:
def __init__(
self,
diff --git a/tests/application/test_routing.py b/tests/application/test_routing.py
index 09b0049a..a1f5dc98 100644
--- a/tests/application/test_routing.py
+++ b/tests/application/test_routing.py
@@ -16,6 +16,51 @@ from free_claude_code.core.openai_responses import OpenAIResponsesRequest
from free_claude_code.core.reasoning import ReasoningControl, ReasoningEffort
+ "prefix,off",
+ [("", False), ("anthropic/", False), ("claude-3-freecc-no-thinking/", True)],
+)
+ "suffix", ["openai/gpt-4.1", "vendor/opus", "sonnet", "haiku", "fable"]
+)
+def test_retired_direct_ids_route_to_default_before_family_aliases(prefix, off, suffix):
+ settings = Settings(
+ MODEL="groq/default",
+ MODEL_OPUS="deepseek/opus",
+ MODEL_SONNET="deepseek/sonnet",
+ MODEL_HAIKU="deepseek/haiku",
+ MODEL_FABLE="deepseek/fable",
+ MODEL_FALLBACKS="groq/default,groq/backup",
+ REASONING_POLICY="high",
+ REASONING_OPUS="off",
+ )
+ original = f"{prefix}github_models/{suffix}"
+ router = ModelRouter(settings)
+ route = router.resolve(original)
+ assert route.original_model == original
+ assert route.primary.provider_model_ref == "groq/default"
+ assert [target.provider_model_ref for target in route.fallbacks] == ["groq/backup"]
+ assert route.reasoning_preference is (
+ ReasoningPreference.OFF if off else ReasoningPreference.HIGH
+ )
+ messages = router.resolve_messages_request(
+ MessagesRequest(
+ model=original,
+ max_tokens=100,
+ messages=[Message(role="user", content="hi")],
+ )
+ )
+ responses = router.resolve_responses_request(
+ OpenAIResponsesRequest(model=original, input="hi")
+ )
+ count = router.resolve_token_count_request(
+ TokenCountRequest(model=original, messages=[Message(role="user", content="hi")])
+ )
+ for routed in (messages, responses, count):
+ assert routed.request.model == "default"
+ assert routed.resolved == route
+
+
@pytest.fixture
def settings():
settings = Settings()
diff --git a/tests/config/test_config.py b/tests/config/test_config.py
index d12edcc6..196e630d 100644
--- a/tests/config/test_config.py
+++ b/tests/config/test_config.py
@@ -11,6 +11,7 @@ from free_claude_code.application.routing import ModelRouter
from free_claude_code.config import loader
from free_claude_code.config.constants import (
ANTHROPIC_DEFAULT_MAX_OUTPUT_TOKENS,
+ DEFAULT_MODEL,
HTTP_CONNECT_TIMEOUT_DEFAULT,
)
from free_claude_code.config.env_files import dotenv_values_from_file
@@ -36,6 +37,82 @@ from free_claude_code.config.reasoning import ReasoningPreference
from free_claude_code.config.settings import Settings
[email protected]("source", [ConfigSource.MANAGED, ConfigSource.PROCESS])
[email protected]("tier", ["FABLE", "OPUS", "SONNET", "HAIKU"])
+def test_retirement_normalizes_sources_without_resurrecting_overrides(source, tier):
+ retired = {
+ "MODEL": "github_models/openai/old",
+ f"MODEL_{tier}": "github_models/vendor/opus",
+ "MODEL_FALLBACKS": "github_models/old",
+ f"REASONING_{tier}": "off",
+ }
+ managed = {
+ "MODEL": "groq/managed",
+ f"MODEL_{tier}": "groq/tier",
+ "MODEL_FALLBACKS": "groq/backup",
+ }
+ snapshot = compose_settings_snapshot(
+ retired if source is ConfigSource.MANAGED else managed,
+ retired if source is ConfigSource.PROCESS else {},
+ )
+ assert snapshot.settings.model == DEFAULT_MODEL
+ assert getattr(snapshot.settings, f"model_{tier.lower()}") is None
+ assert (
+ getattr(snapshot.settings, f"reasoning_{tier.lower()}")
+ is ReasoningPreference.OFF
+ )
+ assert snapshot.settings.model_fallbacks is None
+ expected = (
+ ConfigSource.PROCESS if source is ConfigSource.PROCESS else ConfigSource.DEFAULT
+ )
+ assert snapshot.sources["model"] is expected
+ assert snapshot.sources[f"model_{tier.lower()}"] is expected
+ assert snapshot.sources["model_fallbacks"] is expected
+ assert retired["MODEL"] == "github_models/openai/old"
+
+
+def test_retirement_preserves_effective_default_order_and_process_environment(
+ monkeypatch,
+):
+ monkeypatch.delenv("MODEL", raising=False)
+ monkeypatch.setenv("MODEL_OPUS", "github_models/opus")
+ snapshot = compose_settings_snapshot(
+ {
+ "MODEL": "groq/default",
+ "MODEL_OPUS": "groq/managed-tier",
+ "MODEL_FALLBACKS": "groq/first, github_models/old, deepseek/last",
+ }
+ )
+ assert snapshot.settings.model == "groq/default"
+ assert snapshot.settings.model_opus is None
+ assert snapshot.sources["model_opus"] is ConfigSource.PROCESS
+ assert snapshot.settings.model_fallbacks == ("groq/first", "deepseek/last")
+ assert loader.os.environ["MODEL_OPUS"] == "github_models/opus"
+
+
+ "values",
+ [
+ {"MODEL": "github_models/"},
+ {"MODEL": "other/unknown"},
+ {"MODEL_FALLBACKS": "github_models/old,groq/a,groq/a"},
+ {"MODEL_FALLBACKS": "github_models/old,,groq/a"},
+ {"MODEL_FALLBACKS": "github_models/old,"},
+ {"MODEL_FALLBACKS": ",github_models/old"},
+ {"MODEL_FALLBACKS": "github_models/old, "},
+ {"MODEL_FALLBACKS": "github_models/old,github_models/"},
+ ],
+)
+def test_retirement_does_not_hide_invalid_configuration(values):
+ with pytest.raises(ValidationError):
+ compose_settings_snapshot(values, {})
+
+
+def test_direct_settings_still_reject_retired_provider():
+ with pytest.raises(ValidationError):
+ Settings(MODEL="github_models/openai/old")
+
+
def test_settings_defaults_are_valid_and_nonempty() -> None:
settings = Settings()
diff --git a/tests/config/test_env_migrations.py b/tests/config/test_env_migrations.py
index 630f7863..98c5bfc0 100644
--- a/tests/config/test_env_migrations.py
+++ b/tests/config/test_env_migrations.py
@@ -18,6 +18,66 @@ from free_claude_code.config.env_migrations import (
from free_claude_code.config.loader import resolve_settings_snapshot
[email protected]("schema", ["", "FCC_CONFIG_SCHEMA=1\n"])
+def test_retired_models_repair_managed_config_once_preserving_unknowns(
+ monkeypatch, tmp_path, schema
+):
+ managed, _ = _paths(monkeypatch, tmp_path)
+ managed.parent.mkdir(parents=True)
+ managed.write_text(
+ schema + "MODEL=github_models/old\nMODEL_OPUS=github_models/opus\n"
+ "REASONING_OPUS=off\nMODEL_FALLBACKS=groq/a,github_models/old,deepseek/b\n"
+ "GITHUB_MODELS_TOKEN=retired-secret\nGITHUB_MODELS_PROXY=http://user:[email protected]\n"
+ "UNKNOWN_PRIVATE=unknown-secret\nCUSTOM=\n",
+ encoding="utf-8",
+ )
+ assert consolidate_managed_config({}).changed
+ values = dotenv_values_from_file(managed)
+ assert "MODEL" not in values and "MODEL_OPUS" not in values
+ assert values["MODEL_FALLBACKS"] == "groq/a,deepseek/b"
+ assert values["REASONING_OPUS"] == "off"
+ assert values["GITHUB_MODELS_TOKEN"] == "retired-secret"
+ assert values["GITHUB_MODELS_PROXY"] == "http://user:[email protected]"
+ assert values["CUSTOM"] == ""
+ preview = env_migrations.render_managed_config(values, mask_secrets=True)
+ for secret in ("retired-secret", "password", "unknown-secret"):
+ assert secret not in preview
+ assert "GITHUB_MODELS_TOKEN=********" in preview
+ assert "UNKNOWN_PRIVATE=********" in preview
+ before = managed.read_bytes(), managed.stat().st_mtime_ns
+ assert not consolidate_managed_config({}).changed
+ assert (managed.read_bytes(), managed.stat().st_mtime_ns) == before
+
+
+def test_retired_legacy_import_repairs_refs_without_importing_credentials(
+ monkeypatch, tmp_path
+):
+ managed, legacy = _paths(monkeypatch, tmp_path)
+ legacy.parent.mkdir(parents=True)
+ original = "MODEL=github_models/old\nGITHUB_MODELS_TOKEN=old-secret\nGITHUB_MODELS_PROXY=http://[email protected]\n"
+ legacy.write_text(original, encoding="utf-8")
+ consolidate_managed_config({})
+ values = dotenv_values_from_file(managed)
+ assert not {"MODEL", "GITHUB_MODELS_TOKEN", "GITHUB_MODELS_PROXY"} & values.keys()
+ assert legacy.read_text(encoding="utf-8") == original
+
+
+def test_retirement_atomic_write_failure_preserves_managed_file(monkeypatch, tmp_path):
+ managed, _ = _paths(monkeypatch, tmp_path)
+ managed.parent.mkdir(parents=True)
+ original = b"FCC_CONFIG_SCHEMA=1\nMODEL=github_models/old\n"
+ managed.write_bytes(original)
+
+ def fail_replace(*args):
+ raise OSError("injected replace failure")
+
+ monkeypatch.setattr(env_migrations.os, "replace", fail_replace)
+ with pytest.raises(OSError, match="injected"):
+ consolidate_managed_config({})
+ assert managed.read_bytes() == original
+ assert list(managed.parent.glob("*.tmp")) == []
+
+
def _paths(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> tuple[Path, Path]:
managed = tmp_path / ".fcc" / ".env"
legacy = tmp_path / "legacy" / ".env"
diff --git a/tests/contracts/test_feature_manifest.py b/tests/contracts/test_feature_manifest.py
index cb948606..892323e3 100644
--- a/tests/contracts/test_feature_manifest.py
+++ b/tests/contracts/test_feature_manifest.py
@@ -7,7 +7,6 @@ from free_claude_code.providers.base import BaseProvider
from free_claude_code.providers.cloudflare import CloudflareProvider
from free_claude_code.providers.deepseek import DeepSeekProvider
from free_claude_code.providers.gemini import GeminiProvider
-from free_claude_code.providers.github_models import GitHubModelsProvider
from free_claude_code.providers.groq import GroqProvider
from free_claude_code.providers.kilo import KiloProvider
from free_claude_code.providers.lmstudio import LMStudioProvider
@@ -75,7 +74,6 @@ def test_provider_and_platform_registries_include_builtins() -> None:
"kilo": KiloProvider,
"cloudflare": CloudflareProvider,
"lmstudio": LMStudioProvider,
- "github_models": GitHubModelsProvider,
"groq": GroqProvider,
"gemini": GeminiProvider,
"vertex": VertexProvider,
diff --git a/tests/contracts/test_provider_catalog_order.py b/tests/contracts/test_provider_catalog_order.py
index c419f241..5c6c29f5 100644
--- a/tests/contracts/test_provider_catalog_order.py
+++ b/tests/contracts/test_provider_catalog_order.py
@@ -35,7 +35,6 @@ _EXPECTED_PROVIDER_ORDER: tuple[str, ...] = (
"bedrock",
"huggingface",
"cohere",
- "github_models",
"wafer",
"kimi",
"kimi_code",
diff --git a/tests/contracts/test_smoke_config.py b/tests/contracts/test_smoke_config.py
index c9d0f354..f3e6b203 100644
--- a/tests/contracts/test_smoke_config.py
+++ b/tests/contracts/test_smoke_config.py
@@ -45,7 +45,6 @@ def _settings(**overrides):
"bedrock_base_url": "https://bedrock-mantle.us-east-1.api.aws/v1",
"huggingface_api_key": "",
"cohere_api_key": "",
- "github_models_token": "",
"zai_api_key": "",
"kilo_api_key": "",
"gemini_api_key": "",
@@ -795,22 +794,6 @@ def test_cohere_provider_configuration_uses_api_key(monkeypatch) -> None:
assert models[0].full_model == PROVIDER_SMOKE_DEFAULT_MODELS["cohere"]
-def test_github_models_provider_configuration_uses_token(monkeypatch) -> None:
- monkeypatch.delenv("FCC_SMOKE_MODEL_GITHUB_MODELS", raising=False)
- config = _smoke_config(
- settings=_settings(
- model="ollama/llama3.1",
- ollama_base_url="",
- github_models_token="github-token",
- )
- )
-
- assert config.has_provider_configuration("github_models")
- models = config.provider_smoke_models()
- assert models[0].provider == "github_models"
- assert models[0].full_model == PROVIDER_SMOKE_DEFAULT_MODELS["github_models"]
-
-
def test_sambanova_provider_configuration_uses_api_key(monkeypatch) -> None:
monkeypatch.delenv("FCC_SMOKE_MODEL_SAMBANOVA", raising=False)
config = _smoke_config(
diff --git a/tests/providers/test_credential_validation.py b/tests/providers/test_credential_validation.py
index 50c339cc..9392cc7c 100644
--- a/tests/providers/test_credential_validation.py
+++ b/tests/providers/test_credential_validation.py
@@ -68,7 +68,6 @@ CASES = [
),
("mistral", "https://api.mistral.ai/v1/models", {"data": []}, None),
("wandb", "https://api.inference.wandb.ai/v1/models", {"data": []}, None),
- ("github_models", "https://api.github.com/user", {"login": "user"}, None),
("minimax", "https://api.minimax.io/v1/models", {"data": []}, None),
(
"novita",
diff --git a/tests/providers/test_github_models.py b/tests/providers/test_github_models.py
deleted file mode 100644
index d1777e39..00000000
--- a/tests/providers/test_github_models.py
+++ /dev/null
@@ -1,323 +0,0 @@
-"""Tests for GitHub Models OpenAI-compatible provider."""
-
-from collections.abc import AsyncIterator
-from dataclasses import replace
-from types import SimpleNamespace
-from unittest.mock import AsyncMock, patch
-
-import httpx
-import pytest
-
-from free_claude_code.application.model_metadata import ProviderModelInfo
-from free_claude_code.config.provider_catalog import GITHUB_MODELS_DEFAULT_BASE
-from free_claude_code.core.anthropic.models import Message, MessagesRequest
-from free_claude_code.core.anthropic.stream_contracts import parse_sse_text
-from free_claude_code.providers.base import ProviderConfig
-from free_claude_code.providers.github_models import GitHubModelsProvider
-from free_claude_code.providers.github_models.client import GITHUB_MODELS_CATALOG_URL
-from free_claude_code.providers.model_listing import ModelListResponseError
-from tests.providers.support import (
- REASONING_ON,
- immediate_admission,
- make_provider_config,
-)
-
-
-def github_models_config() -> ProviderConfig:
- return make_provider_config(
- api_key="test-github-models-token",
- base_url=GITHUB_MODELS_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
- )
-
-
-def github_models_provider(
- github_models_config: ProviderConfig,
-) -> GitHubModelsProvider:
- return GitHubModelsProvider(github_models_config, admission=immediate_admission())
-
-
-def _request(model: str = "openai/gpt-4.1") -> MessagesRequest:
- return MessagesRequest(
- model=model,
- max_tokens=100,
- messages=[Message(role="user", content="hi")],
- )
-
-
-def _chunk(delta: SimpleNamespace, *, finish_reason: str = "stop") -> SimpleNamespace:
- return SimpleNamespace(
- choices=[SimpleNamespace(delta=delta, finish_reason=finish_reason)],
- usage=SimpleNamespace(completion_tokens=5, prompt_tokens=8),
- )
-
-
-async def _stream(*chunks: SimpleNamespace) -> AsyncIterator[SimpleNamespace]:
- for chunk in chunks:
- yield chunk
-
-
-def _catalog_response(payload: object) -> httpx.Response:
- return httpx.Response(
- 200,
- json=payload,
- request=httpx.Request("GET", GITHUB_MODELS_CATALOG_URL),
- )
-
-
-def test_default_base_url_constant() -> None:
- assert GITHUB_MODELS_DEFAULT_BASE == "https://models.github.ai/inference"
-
-
-def test_init_uses_default_base_url_api_key_and_github_headers(
- github_models_config: ProviderConfig,
-) -> None:
- with patch(
- "free_claude_code.providers.openai_chat.provider.AsyncOpenAI"
- ) as mock_openai:
- provider = GitHubModelsProvider(
- github_models_config, admission=immediate_admission()
- )
-
- assert provider._api_key == "test-github-models-token"
- assert provider._base_url == GITHUB_MODELS_DEFAULT_BASE
- assert provider._catalog_url == GITHUB_MODELS_CATALOG_URL
- assert mock_openai.call_args.kwargs["base_url"] == GITHUB_MODELS_DEFAULT_BASE
- assert mock_openai.call_args.kwargs["api_key"] == "test-github-models-token"
- assert mock_openai.call_args.kwargs["default_headers"] == {
- "Accept": "application/vnd.github+json",
- "X-GitHub-Api-Version": "2026-03-10",
- }
-
-
-def test_init_strips_trailing_slash(github_models_config: ProviderConfig) -> None:
- config = replace(
- github_models_config,
- base_url=f"{GITHUB_MODELS_DEFAULT_BASE}/",
- )
-
- with patch("free_claude_code.providers.openai_chat.provider.AsyncOpenAI"):
- provider = GitHubModelsProvider(config, admission=immediate_admission())
-
- assert provider._base_url == GITHUB_MODELS_DEFAULT_BASE
-
-
-def test_model_list_headers_use_bearer_auth(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- assert github_models_provider._model_list_headers() == {
- "Accept": "application/vnd.github+json",
- "Authorization": "Bearer test-github-models-token",
- "X-GitHub-Api-Version": "2026-03-10",
- }
-
-
-def test_build_request_body_uses_shared_openai_chat_policy(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- request = _request()
-
- body = github_models_provider._build_request_body(request, reasoning=REASONING_ON)
-
- assert body["model"] == "openai/gpt-4.1"
- assert body["max_tokens"] == 100
- assert "extra_body" not in body
-
-
-async def test_lists_stream_tool_capable_models_only(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- with patch.object(
- github_models_provider._model_list_client,
- "get",
- new_callable=AsyncMock,
- return_value=_catalog_response(
- [
- {
- "id": "openai/gpt-4.1",
- "capabilities": ["streaming", "tool-calling"],
- },
- {
- "id": "openai/text-only",
- "capabilities": ["streaming"],
- },
- {
- "id": "openai/no-stream-tools",
- "capabilities": ["tool-calling"],
- },
- ]
- ),
- ) as mock_get:
- assert await github_models_provider.list_model_infos() == frozenset(
- {ProviderModelInfo("openai/gpt-4.1")}
- )
-
- mock_get.assert_awaited_once_with(
- GITHUB_MODELS_CATALOG_URL,
- headers={
- "Accept": "application/vnd.github+json",
- "Authorization": "Bearer test-github-models-token",
- "X-GitHub-Api-Version": "2026-03-10",
- },
- )
-
-
-async def test_model_list_rejects_malformed_payload(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- with (
- patch.object(
- github_models_provider._model_list_client,
- "get",
- new_callable=AsyncMock,
- return_value=_catalog_response({"data": []}),
- ),
- pytest.raises(ModelListResponseError, match="top-level array"),
- ):
- await github_models_provider.list_model_infos()
-
-
-async def test_model_list_returns_empty_set_when_no_models_support_streaming_tools(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- with patch.object(
- github_models_provider._model_list_client,
- "get",
- new_callable=AsyncMock,
- return_value=_catalog_response(
- [
- {"id": "openai/text-only", "capabilities": ["streaming"]},
- {"id": "openai/non-stream-tool", "capabilities": ["tool-calling"]},
- ]
- ),
- ):
- assert await github_models_provider.list_model_infos() == frozenset()
-
-
-async def test_stream_messages_text(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- delta = SimpleNamespace(
- content="Hello from GitHub Models",
- reasoning_content=None,
- tool_calls=None,
- )
-
- with patch.object(
- github_models_provider._client.chat.completions,
- "create",
- new_callable=AsyncMock,
- return_value=_stream(_chunk(delta)),
- ) as mock_create:
- events = [
- event async for event in github_models_provider.stream_messages(_request())
- ]
-
- parsed = parse_sse_text("".join(events))
- assert any(
- event.event == "content_block_delta"
- and event.data.get("delta", {}).get("text") == "Hello from GitHub Models"
- for event in parsed
- )
- assert mock_create.call_args.kwargs["model"] == "openai/gpt-4.1"
- assert mock_create.call_args.kwargs["stream"] is True
-
-
-async def test_stream_messages_tool_call(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- tool_call = SimpleNamespace(
- index=0,
- id="call_1",
- function=SimpleNamespace(name="echo", arguments='{"value":"x"}'),
- )
- delta = SimpleNamespace(
- content=None, reasoning_content=None, tool_calls=[tool_call]
- )
- request = MessagesRequest.model_validate(
- {
- "model": "openai/gpt-4.1",
- "messages": [{"role": "user", "content": "Use the tool"}],
- "tools": [
- {
- "name": "echo",
- "description": "Echo a value",
- "input_schema": {
- "type": "object",
- "properties": {"value": {"type": "string"}},
- "required": ["value"],
- },
- }
- ],
- }
- )
-
- with patch.object(
- github_models_provider._client.chat.completions,
- "create",
- new_callable=AsyncMock,
- return_value=_stream(_chunk(delta, finish_reason="tool_calls")),
- ):
- events = [
- event async for event in github_models_provider.stream_messages(request)
- ]
-
- parsed = parse_sse_text("".join(events))
- assert any(
- event.event == "content_block_start"
- and event.data.get("content_block", {}).get("type") == "tool_use"
- and event.data.get("content_block", {}).get("name") == "echo"
- for event in parsed
- )
- assert any(
- event.event == "content_block_delta"
- and event.data.get("delta", {}).get("partial_json") == '{"value":"x"}'
- for event in parsed
- )
-
-
-async def test_stream_messages_reasoning_content(
- github_models_provider: GitHubModelsProvider,
-) -> None:
- delta = SimpleNamespace(
- content=None,
- reasoning_content="Thinking via GitHub Models",
- tool_calls=None,
- )
-
- with patch.object(
- github_models_provider._client.chat.completions,
- "create",
- new_callable=AsyncMock,
- return_value=_stream(_chunk(delta)),
- ):
- events = [
- event async for event in github_models_provider.stream_messages(_request())
- ]
-
- parsed = parse_sse_text("".join(events))
- assert any(
- event.event == "content_block_delta"
- and event.data.get("delta", {}).get("thinking") == "Thinking via GitHub Models"
- for event in parsed
- )
-
-
-async def test_cleanup(github_models_provider: GitHubModelsProvider) -> None:
- github_models_provider._client = AsyncMock()
- github_models_provider._model_list_client = AsyncMock()
-
- await github_models_provider.cleanup()
-
- github_models_provider._client.close.assert_called_once()
- github_models_provider._model_list_client.aclose.assert_called_once()
diff --git a/tests/providers/test_provider_runtime.py b/tests/providers/test_provider_runtime.py
index ae8f1068..be49694b 100644
--- a/tests/providers/test_provider_runtime.py
+++ b/tests/providers/test_provider_runtime.py
@@ -18,7 +18,6 @@ from free_claude_code.config.provider_catalog import (
COHERE_DEFAULT_BASE,
DEEPINFRA_DEFAULT_BASE,
FEATHERLESS_DEFAULT_BASE,
- GITHUB_MODELS_DEFAULT_BASE,
HUGGINGFACE_DEFAULT_BASE,
KIMI_CODE_DEFAULT_BASE,
LLM7_DEFAULT_BASE,
@@ -45,7 +44,6 @@ from free_claude_code.providers.admission import ProviderAdmissionController
from free_claude_code.providers.cloudflare import CloudflareProvider
from free_claude_code.providers.deepseek import DeepSeekProvider
from free_claude_code.providers.gemini import GeminiProvider
-from free_claude_code.providers.github_models import GitHubModelsProvider
from free_claude_code.providers.groq import GroqProvider
from free_claude_code.providers.kilo import KiloProvider
from free_claude_code.providers.lmstudio import LMStudioProvider
@@ -97,7 +95,6 @@ def _make_settings(**overrides):
mock.bedrock_base_url = BEDROCK_DEFAULT_BASE
mock.huggingface_api_key = "test_huggingface_key"
mock.cohere_api_key = "test_cohere_key"
- mock.github_models_token = "test_github_models_token"
mock.zai_api_key = "test_zai_key"
mock.tokenrouter_api_key = "test_tokenrouter_key"
mock.tokenrouter_base_url = TOKENROUTER_DEFAULT_BASE
@@ -130,7 +127,6 @@ def _make_settings(**overrides):
mock.bedrock_proxy = None
mock.huggingface_proxy = None
mock.cohere_proxy = None
- mock.github_models_proxy = None
mock.zai_proxy = None
mock.zai_api_proxy = None
mock.tokenrouter_proxy = None
@@ -780,14 +776,6 @@ def test_cohere_descriptor_uses_openai_chat_compatibility_api() -> None:
assert descriptor.proxy_attr == "cohere_proxy"
-def test_github_models_descriptor_uses_openai_chat_inference_api() -> None:
- descriptor = PROVIDER_CATALOG["github_models"]
-
- assert descriptor.default_base_url == GITHUB_MODELS_DEFAULT_BASE
- assert descriptor.credential_env == "GITHUB_MODELS_TOKEN"
- assert descriptor.proxy_attr == "github_models_proxy"
-
-
def test_build_provider_config_vercel_uses_gateway_key_and_proxy() -> None:
descriptor = PROVIDER_CATALOG["vercel"]
settings = _make_settings(
@@ -827,19 +815,6 @@ def test_build_provider_config_cohere_uses_api_key_and_proxy() -> None:
assert config.proxy == "http://proxy.test:8080"
-def test_build_provider_config_github_models_uses_token_and_proxy() -> None:
- descriptor = PROVIDER_CATALOG["github_models"]
- settings = _make_settings(
- github_models_token="github-token",
- github_models_proxy="http://proxy.test:8080",
- )
-
- config = build_provider_config(descriptor, settings)
-
- assert config.api_key == "github-token"
- assert config.proxy == "http://proxy.test:8080"
-
-
def test_create_provider_uses_openai_chat_openrouter_by_default():
with patch("free_claude_code.providers.openai_chat.provider.AsyncOpenAI"):
provider = create_provider("open_router", _make_settings())
@@ -862,7 +837,6 @@ def test_create_provider_instantiates_each_builtin():
bedrock_api_key="test_bedrock_key",
huggingface_api_key="test_huggingface_key",
cohere_api_key="test_cohere_key",
- github_models_token="test_github_models_token",
kimi_api_key="test_kimi_key",
kimi_code_api_key="test_kimi_code_key",
provider_rate_limit=7,
@@ -907,7 +881,6 @@ def test_create_provider_instantiates_each_builtin():
"bedrock": OpenAIChatProvider,
"huggingface": OpenAIChatProvider,
"cohere": OpenAIChatProvider,
- "github_models": GitHubModelsProvider,
"zai": OpenAIChatProvider,
"zai_api": OpenAIChatProvider,
"tokenrouter": OpenAIChatProvider,
diff --git a/tests/runtime/test_chat_sqlite.py b/tests/runtime/test_chat_sqlite.py
index e49c589a..67132c08 100644
--- a/tests/runtime/test_chat_sqlite.py
+++ b/tests/runtime/test_chat_sqlite.py
@@ -19,6 +19,154 @@ from free_claude_code.application.chat import (
from free_claude_code.runtime.chat_sqlite import SQLiteChatStore
[email protected]("disable_reasoning", [True, False])
[email protected]("healthy_preference", [True, False])
+async def test_retirement_repairs_current_selections_only(
+ tmp_path, disable_reasoning, healthy_preference
+):
+ store = _store(tmp_path)
+ await store.start()
+ try:
+ retired = await store.create_session(
+ session_id=_id(),
+ model="github_models/vendor/old",
+ reasoning=ChatReasoning.HIGH,
+ )
+ generation_id = _id()
+ await store.begin_send(
+ retired.id,
+ expected_revision=retired.revision,
+ turn_id=_id(),
+ generation_id=generation_id,
+ operation_id=_id(),
+ user_text="Keep history",
+ requested_model=retired.model,
+ reasoning=retired.reasoning,
+ effective_output_limit=100,
+ )
+ await store.set_generation_actual_model(generation_id, retired.model)
+ await store.finish_generation(
+ generation_id,
+ status=GenerationStatus.COMPLETED,
+ stop_reason="end_turn",
+ error_code=None,
+ error_message=None,
+ )
+ await store.upsert_compaction(
+ retired.id,
+ covered_through_sequence=1,
+ summary="Keep summary",
+ estimated_tokens=5,
+ requested_model=retired.model,
+ actual_model=retired.model,
+ )
+ before = await store.get_transcript(retired.id)
+ untouched = [
+ await store.create_session(
+ session_id=_id(), model=model, reasoning=ChatReasoning.MEDIUM
+ )
+ for model in ("githubXmodels/old", "github_models/", "groq/healthy")
+ ]
+ if not healthy_preference:
+ temporary = await store.create_session(
+ session_id=_id(), model=retired.model, reasoning=ChatReasoning.HIGH
+ )
+ await store.delete_session(
+ temporary.id, expected_revision=temporary.revision
+ )
+ preferences = await store.load_preferences()
+ assert (
+ await store.repair_retired_model_selections(
+ retired_provider_ids=frozenset({"github_models"}),
+ default_model="groq/default",
+ disable_reasoning=disable_reasoning,
+ )
+ == 1
+ )
+ after = await store.get_transcript(retired.id)
+ assert after.session.model == "groq/default"
+ assert after.session.revision == before.session.revision + 1
+ assert after.session.reasoning is (
+ ChatReasoning.OFF if disable_reasoning else ChatReasoning.HIGH
+ )
+ assert after.session.title == before.session.title
+ assert after.turns == before.turns
+ assert after.compaction == before.compaction
+ for session in untouched:
+ assert await store.get_session(session.id) == session
+ updated_preferences = await store.load_preferences()
+ if healthy_preference:
+ assert updated_preferences == preferences
+ else:
+ assert updated_preferences.last_model == "groq/default"
+ assert updated_preferences.last_reasoning is (
+ ChatReasoning.OFF if disable_reasoning else ChatReasoning.HIGH
+ )
+ assert (
+ await store.repair_retired_model_selections(
+ retired_provider_ids=frozenset({"github_models"}),
+ default_model="groq/default",
+ disable_reasoning=disable_reasoning,
+ )
+ == 0
+ )
+ assert await store.get_transcript(retired.id) == after
+ assert await store.load_preferences() == updated_preferences
+ finally:
+ await store.close()
+
+
+async def test_retirement_repairs_global_preference_without_sessions(tmp_path):
+ store = _store(tmp_path)
+ await store.start()
+ try:
+ session = await store.create_session(
+ session_id=_id(), model="github_models/old", reasoning=ChatReasoning.HIGH
+ )
+ await store.delete_session(session.id, expected_revision=session.revision)
+ assert (
+ await store.repair_retired_model_selections(
+ retired_provider_ids=frozenset({"github_models"}),
+ default_model="groq/default",
+ disable_reasoning=True,
+ )
+ == 0
+ )
+ preferences = await store.load_preferences()
+ assert preferences.last_model == "groq/default"
+ assert preferences.last_reasoning is ChatReasoning.OFF
+ finally:
+ await store.close()
+
+
+async def test_retirement_rolls_back_sessions_when_preference_update_fails(tmp_path):
+ store = _store(tmp_path)
+ await store.start()
+ try:
+ session = await store.create_session(
+ session_id=_id(), model="github_models/old", reasoning=ChatReasoning.HIGH
+ )
+ before = await store.load_preferences()
+ with closing(sqlite3.connect(tmp_path / "chat.db")) as connection:
+ connection.execute(
+ "CREATE TRIGGER fail_preference BEFORE UPDATE ON chat_settings BEGIN SELECT RAISE(ABORT, 'injected'); END"
+ )
+ connection.commit()
+ with pytest.raises(ChatUnavailableError):
+ await store.repair_retired_model_selections(
+ retired_provider_ids=frozenset({"github_models"}),
+ default_model="groq/default",
+ disable_reasoning=True,
+ )
+ assert await store.get_session(session.id) == session
+ assert await store.load_preferences() == before
+ finally:
+ await store.close()
+
+
def _id() -> str:
return str(uuid.uuid4())
diff --git a/uv.lock b/uv.lock
index 73ad1cb7..3758cfa3 100644
--- a/uv.lock
+++ b/uv.lock
@@ -598,7 +598,7 @@ wheels = [
[[package]]
name = "free-claude-code"
-version = "5.20.0"
+version = "5.21.0"
source = { editable = "." }
dependencies = [
{ name = "aiohttp" },