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

Add Cheaper Inference provider (#1899)

aiapienthusiast4340ffc2384f
diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
index ca058887..8ef74adc 100644
--- a/.github/workflows/tests.yml
+++ b/.github/workflows/tests.yml
@@ -173,6 +173,7 @@ jobs:
- name: Run Playwright
run: >-
uv run --locked pytest e2e -v --tb=short
+ -o faulthandler_timeout=10 --durations=20
--tracing=retain-on-failure --screenshot=only-on-failure
--full-page-screenshot --output=test-results
diff --git a/README.md b/README.md
index 8125c497..cd41bed3 100644
--- a/README.md
+++ b/README.md
@@ -25,7 +25,7 @@
## What You Get
-- **54 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.
+- **55 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. This works 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.
@@ -253,6 +253,7 @@ from more than one provider before succeeding.
| [Scaleway](https://console.scaleway.com/iam/api-keys) | `SCW_SECRET_KEY` | `scaleway/deepseek/deepseek-v4-flash` |
| [Lightning AI](https://lightning.ai/) | `LIGHTNING_API_KEY` | `lightning/lightning-ai/Qwen3.8-27B` |
| [Experiential Labs](https://platform.experientiallabs.ai/) | `EXPLABS_API_KEY` | `experiential/union-alpha` |
+| [Cheaper Inference](https://cheaperinference.com/signup) | `CHEAPER_INFERENCE_API_KEY` | `cheaperinference/gpt-5.4-mini` |
| [Ollama Cloud](https://ollama.com/settings/keys) | `OLLAMA_API_KEY` | `ollama_cloud/qwen3-coder:480b` |
| [LM Studio](https://lmstudio.ai/) | `LM_STUDIO_BASE_URL` | `lmstudio/<model-id>` |
| [llama.cpp](https://github.com/ggml-org/llama.cpp) | `LLAMACPP_BASE_URL` | `llamacpp/<model-id>` |
diff --git a/e2e/conftest.py b/e2e/conftest.py
index ece94712..5d7bdff6 100644
--- a/e2e/conftest.py
+++ b/e2e/conftest.py
@@ -1,6 +1,7 @@
"""Isolated browser-test composition for the local Admin UI."""
import asyncio
+import faulthandler
import socket
import sys
import threading
@@ -348,6 +349,7 @@ def admin_base_url(
listener.close()
clear_settings_cache()
if thread.is_alive():
+ faulthandler.dump_traceback(file=sys.stderr)
pytest.fail("Admin browser-test server did not stop")
diff --git a/e2e/test_connected_accounts.py b/e2e/test_connected_accounts.py
index f1417400..a5fdcef1 100644
--- a/e2e/test_connected_accounts.py
+++ b/e2e/test_connected_accounts.py
@@ -1,5 +1,7 @@
"""Rendered connected-account flows use provider-owned capabilities and identity."""
+import time
+
import pytest
from playwright.sync_api import Dialog, Page, Route, expect
@@ -55,6 +57,15 @@ class _Accounts:
self.hold_login = False
self.pending_login: list[Route] = []
+ def take_pending_status(self, page: Page) -> Route:
+ deadline = time.monotonic() + 5
+ while not self.pending_status:
+ if time.monotonic() >= deadline:
+ pytest.fail("Account status request did not reach the route handler")
+ # Pump Playwright events so the route handler can enqueue the request.
+ page.wait_for_timeout(10)
+ return self.pending_status.pop(0)
+
def config(self, route: Route) -> None:
response = route.fetch()
config = response.json()
@@ -163,9 +174,6 @@ def _open(page: Page, admin_base_url: str) -> None:
def test_account_modes_wait_for_status_and_recover_after_load_failure(
page: Page, admin_base_url: str, accounts: _Accounts
) -> None:
- page.expose_function(
- "pendingAccountStatusCount", lambda: len(accounts.pending_status)
- )
accounts.hold_status.add("github_copilot")
page.goto(f"{admin_base_url}/admin")
copilot = page.locator('[data-provider="github_copilot"]')
@@ -173,9 +181,7 @@ def test_account_modes_wait_for_status_and_recover_after_load_failure(
expect(copilot.get_by_role("button", name="Loading…", exact=True)).to_be_disabled()
expect(copilot.get_by_role("button", name="Connect", exact=True)).to_have_count(0)
expect(openai.get_by_role("button", name="Connect", exact=True)).to_be_enabled()
- page.wait_for_function("async () => await window.pendingAccountStatusCount() === 1")
-
- accounts.pending_status.pop().fulfill(
+ accounts.take_pending_status(page).fulfill(
status=503, json={"detail": "Account status unavailable."}
)
expect(copilot.locator(".provider-meta")).to_have_text(
@@ -184,9 +190,9 @@ def test_account_modes_wait_for_status_and_recover_after_load_failure(
expect(copilot.get_by_role("button", name="Connect", exact=True)).to_have_count(0)
copilot.get_by_role("button", name="Retry", exact=True).click()
expect(copilot.get_by_role("button", name="Loading…", exact=True)).to_be_disabled()
- page.wait_for_function("async () => await window.pendingAccountStatusCount() === 1")
+ retry = accounts.take_pending_status(page)
accounts.hold_status.clear()
- accounts.pending_status.pop().fulfill(json=accounts.statuses["github_copilot"])
+ retry.fulfill(json=accounts.statuses["github_copilot"])
expect(copilot.get_by_role("button", name="Connect", exact=True)).to_be_enabled()
expect(copilot.get_by_role("button", name="Use device code")).to_have_count(0)
diff --git a/pyproject.toml b/pyproject.toml
index beb914df..4e62a6cd 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
[project]
name = "free-claude-code"
-version = "6.2.65"
+version = "6.2.66"
description = "Local proxy connecting coding agents to OpenAI-compatible AI providers"
readme = "README.md"
license = "AGPL-3.0-only"
diff --git a/smoke/lib/config.py b/smoke/lib/config.py
index 448c154c..594d52f1 100644
--- a/smoke/lib/config.py
+++ b/smoke/lib/config.py
@@ -95,6 +95,7 @@ PROVIDER_SMOKE_DEFAULT_MODELS: dict[str, str] = {
"llm7": "llm7/default",
"lightning": "lightning/lightning-ai/Qwen3.8-27B",
"experiential": "experiential/union-alpha",
+ "cheaperinference": "cheaperinference/gpt-5.4-mini",
"agnes": "agnes/agnes-2.0-flash",
"zenmux": "zenmux/deepseek/deepseek-v4-flash-free",
"wandb": "wandb/openai/gpt-oss-20b",
diff --git a/src/free_claude_code/api/admin_static/providers/SOURCES.md b/src/free_claude_code/api/admin_static/providers/SOURCES.md
index 88e08b52..e8fb62f0 100644
--- a/src/free_claude_code/api/admin_static/providers/SOURCES.md
+++ b/src/free_claude_code/api/admin_static/providers/SOURCES.md
@@ -11,6 +11,7 @@ The LLaMA.cpp icon comes from the upstream llama.cpp repository; its MIT license
| `azure-color.svg` | https://unpkg.com/@lobehub/[email protected]/icons/azure-color.svg |
| `bedrock-color.svg` | https://unpkg.com/@lobehub/[email protected]/icons/bedrock-color.svg |
| `cerebras-color.svg` | https://unpkg.com/@lobehub/[email protected]/icons/cerebras-color.svg |
+| `cheaperinference.svg` | https://www.cheaperinference.com/icon.svg |
| `chutes.png` | https://chutes.ai/favicon.png |
| `cline.svg` | https://unpkg.com/@lobehub/[email protected]/icons/cline.svg |
| `cloudflare-color.svg` | https://unpkg.com/@lobehub/[email protected]/icons/cloudflare-color.svg |
diff --git a/src/free_claude_code/api/admin_static/providers/cheaperinference.svg b/src/free_claude_code/api/admin_static/providers/cheaperinference.svg
new file mode 100644
index 00000000..f389b7a4
--- /dev/null
+++ b/src/free_claude_code/api/admin_static/providers/cheaperinference.svg
@@ -0,0 +1,5 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64">
+ <rect x=".75" y=".75" width="62.5" height="62.5" rx="13.5" fill="#fff" stroke="#040d10" stroke-opacity=".09"/>
+ <path fill="#040d10" d="M38.25 15.75H27.4C18.07 15.75 10.5 23.03 10.5 32c0 8.97 7.52 16.25 16.85 16.25l10.9-10.7h-9.7a5.57 5.57 0 0 1-5.6-5.55c0-3.1 2.5-5.6 5.6-5.6h9.7V15.75Z"/>
+ <path fill="#040d10" d="m41.4 25.55 7.74-8.1c1.48-1.55 4.11-.5 4.11 1.64v26.59a2.57 2.57 0 0 1-2.57 2.57H41.4v-22.7Z"/>
+</svg>
diff --git a/src/free_claude_code/config/admin/provider_manifest.py b/src/free_claude_code/config/admin/provider_manifest.py
index f79999bd..404e926f 100644
--- a/src/free_claude_code/config/admin/provider_manifest.py
+++ b/src/free_claude_code/config/admin/provider_manifest.py
@@ -297,6 +297,14 @@ _PROVIDER_FIELD_OVERRIDES: dict[str, ProviderFieldOverride] = {
"hex characters; mint one at platform.experientiallabs.ai/settings/api-keys."
),
},
+ "CHEAPER_INFERENCE_API_KEY": {
+ "label": "Cheaper Inference API Key",
+ "description": (
+ "Cheaper Inference OpenAI-compatible gateway API key for "
+ "api.cheaperinference.com/v1. "
+ "Keys start with ci_live_; create one at cheaperinference.com/signup."
+ ),
+ },
}
diff --git a/src/free_claude_code/config/provider_catalog.py b/src/free_claude_code/config/provider_catalog.py
index 90f7aef9..d49f31ab 100644
--- a/src/free_claude_code/config/provider_catalog.py
+++ b/src/free_claude_code/config/provider_catalog.py
@@ -86,6 +86,8 @@ LLM7_DEFAULT_BASE = "https://api.llm7.io/v1"
LIGHTNING_DEFAULT_BASE = "https://lightning.ai/api/v1"
# Experiential Labs OpenAI-compatible Chat Completions gateway.
EXPERIENTIAL_DEFAULT_BASE = "https://api.experientiallabs.ai/v1"
+# Cheaper Inference OpenAI-compatible Chat Completions gateway.
+CHEAPERINFERENCE_DEFAULT_BASE = "https://api.cheaperinference.com/v1"
# Agnes AI OpenAI-compatible Chat Completions API.
AGNES_DEFAULT_BASE = "https://apihub.agnes-ai.com/v1"
# ZenMux OpenAI-compatible Chat Completions gateway.
@@ -687,6 +689,17 @@ PROVIDER_CATALOG: dict[str, ProviderDescriptor] = {
default_base_url=EXPERIENTIAL_DEFAULT_BASE,
proxy_attr="experiential_proxy",
),
+ "cheaperinference": ProviderDescriptor(
+ provider_id="cheaperinference",
+ display_name="Cheaper Inference",
+ website_url="https://cheaperinference.com/",
+ logo_filename="cheaperinference.svg",
+ credential_env="CHEAPER_INFERENCE_API_KEY",
+ credential_url="https://cheaperinference.com/signup",
+ credential_attr="cheaperinference_api_key",
+ default_base_url=CHEAPERINFERENCE_DEFAULT_BASE,
+ proxy_attr="cheaperinference_proxy",
+ ),
"ollama_cloud": ProviderDescriptor(
provider_id="ollama_cloud",
display_name="Ollama Cloud",
diff --git a/src/free_claude_code/config/settings.py b/src/free_claude_code/config/settings.py
index 0434b02a..a0bf7111 100644
--- a/src/free_claude_code/config/settings.py
+++ b/src/free_claude_code/config/settings.py
@@ -195,6 +195,11 @@ class Settings(BaseModel):
default=None, validation_alias="EXPLABS_API_KEY"
)
+ # ==================== Cheaper Inference (OpenAI-compatible) ====================
+ cheaperinference_api_key: OptionalNonEmptyString = Field(
+ default=None, validation_alias="CHEAPER_INFERENCE_API_KEY"
+ )
+
# ==================== Fireworks AI Config ====================
fireworks_api_key: OptionalNonEmptyString = Field(
default=None, validation_alias="FIREWORKS_API_KEY"
@@ -498,6 +503,9 @@ class Settings(BaseModel):
experiential_proxy: OptionalNonEmptyString = Field(
default=None, validation_alias="EXPLABS_PROXY"
)
+ cheaperinference_proxy: OptionalNonEmptyString = Field(
+ default=None, validation_alias="CHEAPER_INFERENCE_PROXY"
+ )
fireworks_proxy: OptionalNonEmptyString = Field(
default=None, validation_alias="FIREWORKS_PROXY"
)
diff --git a/src/free_claude_code/providers/credential_validation.py b/src/free_claude_code/providers/credential_validation.py
index ac9fe823..01b489bc 100644
--- a/src/free_claude_code/providers/credential_validation.py
+++ b/src/free_claude_code/providers/credential_validation.py
@@ -156,6 +156,7 @@ _PROBES = (
# https://inference-docs.cerebras.ai/api-reference/models/list-models
# https://docs.sambanova.ai/docs/api-reference/models/get-environments-available-model-list-metadata
# https://docs.fireworks.ai/api-reference/list-accounts
+ # https://cheaperinference.com/docs
_Probe("deepinfra", "https://api.deepinfra.com/v1/me", _identity("uid")),
_Probe("mistral", "/models", _MODELS),
_Probe("wandb", "/models", _MODELS),
@@ -167,6 +168,7 @@ _PROBES = (
),
_Probe("cerebras", "/models", _MODELS),
_Probe("sambanova", "/models", _MODELS),
+ _Probe("cheaperinference", "/models", _MODELS),
_Probe(
"fireworks",
"https://api.fireworks.ai/v1/accounts?pageSize=1",
diff --git a/src/free_claude_code/providers/openai_chat/profiles.py b/src/free_claude_code/providers/openai_chat/profiles.py
index d6ad63df..7ba9702f 100644
--- a/src/free_claude_code/providers/openai_chat/profiles.py
+++ b/src/free_claude_code/providers/openai_chat/profiles.py
@@ -723,6 +723,24 @@ OPENAI_CHAT_PROFILES: dict[str, OpenAIChatProfile] = {
),
model_listing=OpenAIModelListing(path="/models"),
),
+ "cheaperinference": OpenAIChatProfile(
+ _policy(
+ "CHEAPERINFERENCE",
+ ReasoningReplayMode.REASONING_CONTENT,
+ default_max_tokens=ANTHROPIC_DEFAULT_MAX_OUTPUT_TOKENS,
+ ),
+ NamedEffortReasoning(
+ _LOW_MEDIUM_HIGH,
+ enabled_value="medium",
+ ),
+ model_listing=OpenAIModelListing(
+ path="/models",
+ query_params=(("type", "text"), ("streaming", "true")),
+ required_path_values=((("type",), ("text",)),),
+ context_window_tokens_path=("context_length",),
+ max_output_tokens_path=("max_output_tokens",),
+ ),
+ ),
"ollama_cloud": OpenAIChatProfile(
_policy(
"OLLAMA_CLOUD",
diff --git a/tests/contracts/test_provider_catalog_order.py b/tests/contracts/test_provider_catalog_order.py
index 485fbb6f..826dabed 100644
--- a/tests/contracts/test_provider_catalog_order.py
+++ b/tests/contracts/test_provider_catalog_order.py
@@ -56,6 +56,7 @@ _EXPECTED_PROVIDER_ORDER: tuple[str, ...] = (
"scaleway",
"lightning",
"experiential",
+ "cheaperinference",
"ollama_cloud",
"lmstudio",
"llamacpp",
diff --git a/tests/contracts/test_smoke_config.py b/tests/contracts/test_smoke_config.py
index 4a49fb76..9457c910 100644
--- a/tests/contracts/test_smoke_config.py
+++ b/tests/contracts/test_smoke_config.py
@@ -70,6 +70,7 @@ def _settings(**overrides):
"llm7_api_key": "",
"lightning_api_key": "",
"experiential_api_key": "",
+ "cheaperinference_api_key": "",
"fireworks_api_key": "",
"novita_api_key": "",
"cloudflare_api_token": "",
@@ -337,6 +338,58 @@ def test_experiential_is_not_enabled_without_explicit_credential(
assert config.provider_smoke_models() == []
+def test_cheaperinference_provider_configuration_uses_default_model(
+ monkeypatch,
+) -> None:
+ monkeypatch.delenv("FCC_SMOKE_MODEL_CHEAPERINFERENCE", raising=False)
+ config = _smoke_config(
+ settings=_settings(
+ model="ollama/llama3.1",
+ ollama_base_url="",
+ cheaperinference_api_key="ci_live_key",
+ )
+ )
+
+ assert config.has_provider_configuration("cheaperinference")
+ models = config.provider_smoke_models()
+ assert [model.provider for model in models] == ["cheaperinference"]
+ assert models[0].full_model == "cheaperinference/gpt-5.4-mini"
+ assert models[0].source == "provider_default"
+
+
+def test_cheaperinference_smoke_override_accepts_model_with_or_without_prefix(
+ monkeypatch,
+) -> None:
+ settings = _settings(
+ model="ollama/llama3.1",
+ ollama_base_url="",
+ cheaperinference_api_key="ci_live_key",
+ )
+ for override in (
+ "claude-sonnet-5",
+ "cheaperinference/claude-sonnet-5",
+ ):
+ monkeypatch.setenv("FCC_SMOKE_MODEL_CHEAPERINFERENCE", override)
+ models = _smoke_config(settings=settings).provider_smoke_models()
+
+ assert [model.provider for model in models] == ["cheaperinference"]
+ assert models[0].full_model == "cheaperinference/claude-sonnet-5"
+ assert models[0].source == "FCC_SMOKE_MODEL_CHEAPERINFERENCE"
+
+
+def test_cheaperinference_is_not_enabled_without_explicit_credential(
+ monkeypatch,
+) -> None:
+ monkeypatch.delenv("FCC_SMOKE_MODEL_CHEAPERINFERENCE", raising=False)
+ config = _smoke_config(
+ provider_matrix=frozenset({"cheaperinference"}),
+ settings=_settings(ollama_base_url="", cheaperinference_api_key=""),
+ )
+
+ assert not config.has_provider_configuration("cheaperinference")
+ assert config.provider_smoke_models() == []
+
+
def test_xai_provider_smoke_uses_current_grok_model(monkeypatch) -> None:
monkeypatch.delenv("FCC_SMOKE_MODEL_XAI", raising=False)
config = _smoke_config(
diff --git a/tests/providers/test_cheaperinference.py b/tests/providers/test_cheaperinference.py
new file mode 100644
index 00000000..bf71d9d4
--- /dev/null
+++ b/tests/providers/test_cheaperinference.py
@@ -0,0 +1,273 @@
+"""Tests for the Cheaper Inference OpenAI-chat provider profile."""
+
+from unittest.mock import AsyncMock
+
+import httpx2
+import pytest
+from openai import AsyncOpenAI
+
+from free_claude_code.application.model_metadata import ProviderModelInfo
+from free_claude_code.config.constants import ANTHROPIC_DEFAULT_MAX_OUTPUT_TOKENS
+from free_claude_code.config.provider_catalog import CHEAPERINFERENCE_DEFAULT_BASE
+from free_claude_code.core.anthropic.models import MessagesRequest
+from free_claude_code.core.json_types import JsonObject, JsonValue
+from free_claude_code.core.reasoning import ReasoningEffort, ReasoningPolicy
+from free_claude_code.providers.model_listing import ModelListResponseError
+from free_claude_code.providers.openai_chat import OpenAIChatProvider
+from free_claude_code.providers.openai_chat.profiles import OPENAI_CHAT_PROFILES
+from tests.providers.support import (
+ immediate_admission,
+ make_provider_config,
+ profiled_provider,
+ reasoning_for,
+)
+
+_MODEL = "gpt-5.4-mini"
+
+
+def cheaperinference_provider() -> OpenAIChatProvider:
+ return profiled_provider(
+ "cheaperinference",
+ make_provider_config(
+ api_key="test-cheaperinference-key",
+ base_url=CHEAPERINFERENCE_DEFAULT_BASE,
+ rate_limit=10,
+ rate_window=60,
+ ),
+ admission=immediate_admission(provider_name="cheaperinference", max_attempts=1),
+ )
+
+
+def _request(**overrides: JsonValue) -> MessagesRequest:
+ payload: JsonObject = {
+ "model": _MODEL,
+ "messages": [{"role": "user", "content": "Inspect the file."}],
+ "tools": [
+ {
+ "name": "read_file",
+ "description": "Read a file",
+ "input_schema": {
+ "type": "object",
+ "properties": {"path": {"type": "string"}},
+ "required": ["path"],
+ },
+ }
+ ],
+ }
+ payload.update(overrides)
+ return MessagesRequest.model_validate(payload)
+
+
+def test_constructs_standard_openai_chat_provider(
+ cheaperinference_provider: OpenAIChatProvider,
+) -> None:
+ assert isinstance(cheaperinference_provider, OpenAIChatProvider)
+ assert cheaperinference_provider._provider_name == "CHEAPERINFERENCE"
+ assert cheaperinference_provider._api_key == "test-cheaperinference-key"
+ assert cheaperinference_provider._base_url == CHEAPERINFERENCE_DEFAULT_BASE
+
+
+ "reasoning",
+ [ReasoningPolicy.provider_default(), ReasoningPolicy.off()],
+)
+def test_default_and_off_reasoning_omit_effort_and_keep_standard_fields(
+ cheaperinference_provider: OpenAIChatProvider,
+ reasoning: ReasoningPolicy,
+) -> None:
+ body = cheaperinference_provider._chat._build_request_body(
+ _request(), reasoning=reasoning
+ )
+
+ assert body["max_tokens"] == ANTHROPIC_DEFAULT_MAX_OUTPUT_TOKENS
+ assert body["model"] == _MODEL
+ assert body["tools"][0]["function"]["name"] == "read_file"
+ assert "reasoning" not in body
+ assert "reasoning_effort" not in body
+ assert "thinking" not in body
+ assert "extra_body" not in body
+
+
+ ("reasoning", "expected"),
+ (
+ (ReasoningPolicy.on(), "medium"),
+ (ReasoningPolicy.on(effort=ReasoningEffort.MINIMAL), "low"),
+ (ReasoningPolicy.on(effort=ReasoningEffort.LOW), "low"),
+ (ReasoningPolicy.on(effort=ReasoningEffort.MEDIUM), "medium"),
+ (ReasoningPolicy.on(effort=ReasoningEffort.HIGH), "high"),
+ (ReasoningPolicy.on(effort=ReasoningEffort.XHIGH), "high"),
+ (ReasoningPolicy.on(effort=ReasoningEffort.MAX), "high"),
+ ),
+)
+def test_reasoning_maps_efforts_onto_documented_vocabulary(
+ cheaperinference_provider: OpenAIChatProvider,
+ reasoning: ReasoningPolicy,
+ expected: str,
+) -> None:
+ body = cheaperinference_provider._chat._build_request_body(
+ _request(), reasoning=reasoning
+ )
+
+ assert body["reasoning_effort"] == expected
+
+
+def test_replays_reasoning_content_with_tool_history(
+ cheaperinference_provider: OpenAIChatProvider,
+) -> None:
+ request = _request(
+ messages=[
+ {"role": "user", "content": "Inspect the file."},
+ {
+ "role": "assistant",
+ "content": [
+ {"type": "thinking", "thinking": "Read it first."},
+ {"type": "text", "text": "I will inspect it."},
+ {
+ "type": "tool_use",
+ "id": "toolu_1",
+ "name": "read_file",
+ "input": {"path": "example.py"},
+ },
+ ],
+ },
+ {
+ "role": "user",
+ "content": [
+ {
+ "type": "tool_result",
+ "tool_use_id": "toolu_1",
+ "content": "print('hello')",
+ }
+ ],
+ },
+ ]
+ )
+
+ body = cheaperinference_provider._chat._build_request_body(
+ request,
+ reasoning=reasoning_for(request),
+ )
+
+ assert body["messages"][1] == {
+ "role": "assistant",
+ "content": "I will inspect it.",
+ "reasoning_content": "Read it first.",
+ "tool_calls": [
+ {
+ "id": "toolu_1",
+ "type": "function",
+ "function": {
+ "name": "read_file",
+ "arguments": '{"path": "example.py"}',
+ },
+ }
+ ],
+ }
+ assert body["messages"][2] == {
+ "role": "tool",
+ "tool_call_id": "toolu_1",
+ "content": "print('hello')",
+ }
+
+
+async def test_model_catalog_keeps_text_models_with_token_limits(
+ cheaperinference_provider: OpenAIChatProvider,
+) -> None:
+ cheaperinference_provider._client.get = AsyncMock(
+ return_value={
+ "data": [
+ {
+ "id": _MODEL,
+ "type": "text",
+ "context_length": 400_000,
+ "max_output_tokens": 128_000,
+ },
+ {
+ "id": "claude-sonnet-5",
+ "type": "text",
+ "context_length": 1_000_000,
+ "max_output_tokens": 64_000,
+ },
+ {"id": "image-model", "type": "image"},
+ {"id": "video-model", "type": "video"},
+ ]
+ }
+ )
+
+ assert await cheaperinference_provider.list_model_infos() == frozenset(
+ {
+ ProviderModelInfo(
+ _MODEL,
+ context_window_tokens=400_000,
+ max_output_tokens=128_000,
+ ),
+ ProviderModelInfo(
+ "claude-sonnet-5",
+ context_window_tokens=1_000_000,
+ max_output_tokens=64_000,
+ ),
+ }
+ )
+
+
+async def test_model_catalog_rejects_missing_id(
+ cheaperinference_provider: OpenAIChatProvider,
+) -> None:
+ cheaperinference_provider._client.get = AsyncMock(
+ return_value={"data": [{"type": "text"}]}
+ )
+
+ with pytest.raises(ModelListResponseError, match="include id"):
+ await cheaperinference_provider.list_model_infos()
+
+
+async def test_model_catalog_rejects_empty_catalog(
+ cheaperinference_provider: OpenAIChatProvider,
+) -> None:
+ cheaperinference_provider._client.get = AsyncMock(return_value={"data": []})
+
+ with pytest.raises(ModelListResponseError, match="did not include any model ids"):
+ await cheaperinference_provider.list_model_infos()
+
+
+async def test_model_catalog_requests_streaming_text_models_with_bearer_auth() -> None:
+ requests: list[httpx2.Request] = []
+
+ def handler(request: httpx2.Request) -> httpx2.Response:
+ requests.append(request)
+ models = [{"id": _MODEL, "type": "text"}]
+ if request.url.params.get("streaming") != "true":
+ models.append({"id": "gpt-5.5-pro", "type": "text"})
+ if request.url.params.get("type") != "text":
+ models.append({"id": "image-model", "type": "image"})
+ return httpx2.Response(200, json={"data": models})
+
+ async with AsyncOpenAI(
+ api_key="wire-cheaperinference-key",
+ base_url=CHEAPERINFERENCE_DEFAULT_BASE,
+ max_retries=0,
+ http_client=httpx2.AsyncClient(transport=httpx2.MockTransport(handler)),
+ ) as client:
+ cheaperinference_provider = OpenAIChatProvider(
+ make_provider_config(
+ api_key="wire-cheaperinference-key",
+ base_url=CHEAPERINFERENCE_DEFAULT_BASE,
+ ),
+ profile=OPENAI_CHAT_PROFILES["cheaperinference"],
+ admission=immediate_admission(provider_name="cheaperinference"),
+ client=client,
+ )
+ model_infos = await cheaperinference_provider.list_model_infos()
+
+ assert model_infos == frozenset({ProviderModelInfo(_MODEL)})
+ assert len(requests) == 1
+ assert str(requests[0].url) == (
+ "https://api.cheaperinference.com/v1/models?type=text&streaming=true"
+ )
+ assert requests[0].headers["authorization"] == "Bearer wire-cheaperinference-key"
diff --git a/tests/providers/test_credential_validation.py b/tests/providers/test_credential_validation.py
index be92287d..8db66e8d 100644
--- a/tests/providers/test_credential_validation.py
+++ b/tests/providers/test_credential_validation.py
@@ -79,6 +79,12 @@ CASES = [
),
("cerebras", "https://api.cerebras.ai/v1/models", {"data": []}, None),
("sambanova", "https://api.sambanova.ai/v1/models", {"data": []}, None),
+ (
+ "cheaperinference",
+ "https://api.cheaperinference.com/v1/models",
+ {"data": []},
+ None,
+ ),
(
"fireworks",
"https://api.fireworks.ai/v1/accounts?pageSize=1",
diff --git a/tests/providers/test_provider_runtime.py b/tests/providers/test_provider_runtime.py
index fb443fa4..78e525ae 100644
--- a/tests/providers/test_provider_runtime.py
+++ b/tests/providers/test_provider_runtime.py
@@ -14,6 +14,7 @@ from free_claude_code.config.nim import NimSettings
from free_claude_code.config.provider_catalog import (
AGNES_DEFAULT_BASE,
BEDROCK_DEFAULT_BASE,
+ CHEAPERINFERENCE_DEFAULT_BASE,
CHUTES_DEFAULT_BASE,
CLINE_DEFAULT_BASE,
COHERE_DEFAULT_BASE,
@@ -114,6 +115,7 @@ def _make_settings(**overrides):
mock.llm7_api_key = "test_llm7_key"
mock.lightning_api_key = "test_lightning_key"
mock.experiential_api_key = "test_experiential_key"
+ mock.cheaperinference_api_key = "test_cheaperinference_key"
mock.nvidia_nim_proxy = None
mock.open_router_proxy = None
mock.lmstudio_proxy = None
@@ -162,6 +164,7 @@ def _make_settings(**overrides):
mock.llm7_proxy = None
mock.lightning_proxy = None
mock.experiential_proxy = None
+ mock.cheaperinference_proxy = None
mock.kilo_api_key = "test_kilo_key"
mock.kilo_proxy = None
mock.openai_proxy = None
@@ -323,6 +326,31 @@ async def test_experiential_provider_config_uses_key_base_and_proxy() -> None:
assert isinstance(provider, OpenAIChatProvider)
+async def test_cheaperinference_provider_config_uses_key_base_and_proxy() -> None:
+ descriptor = PROVIDER_CATALOG["cheaperinference"]
+ settings = _make_settings(
+ cheaperinference_api_key="ci_live_token",
+ cheaperinference_proxy="http://proxy.test:8080",
+ )
+
+ config = build_provider_config(descriptor, settings)
+ with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
+ provider = await create_provider("cheaperinference", settings)
+
+ assert descriptor.display_name == "Cheaper Inference"
+ assert descriptor.credential_env == "CHEAPER_INFERENCE_API_KEY"
+ assert descriptor.credential_attr == "cheaperinference_api_key"
+ assert descriptor.credential_url == "https://cheaperinference.com/signup"
+ assert descriptor.default_base_url == CHEAPERINFERENCE_DEFAULT_BASE
+ assert descriptor.base_url_attr is None
+ assert descriptor.proxy_attr == "cheaperinference_proxy"
+ assert config.api_key == "ci_live_token"
+ assert config.base_url == "https://api.cheaperinference.com/v1"
+ assert config.proxy == "http://proxy.test:8080"
+ assert isinstance(provider, OpenAIChatProvider)
+
+
@pytest.mark.asyncio
async def test_xai_provider_config_uses_key_base_and_proxy() -> None:
descriptor = PROVIDER_CATALOG["xai"]
@@ -962,6 +990,7 @@ async def test_create_provider_instantiates_each_builtin():
"llm7": OpenAIChatProvider,
"lightning": OpenAIChatProvider,
"experiential": OpenAIChatProvider,
+ "cheaperinference": OpenAIChatProvider,
"opencode_go": OpenCodeProvider,
"vercel": OpenAIChatProvider,
"bedrock": OpenAIChatProvider,
diff --git a/uv.lock b/uv.lock
index 9183db8a..489a472d 100644
--- a/uv.lock
+++ b/uv.lock
@@ -599,7 +599,7 @@ wheels = [
[[package]]
name = "free-claude-code"
-version = "6.2.65"
+version = "6.2.66"
source = { editable = "." }
dependencies = [
{ name = "aiohttp" },