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

patch: Filter unavailable NVIDIA hosted models (#2071)

Ali Khokharcb0c77778f99
diff --git a/src/free_claude_code/providers/nvidia_nim/client.py b/src/free_claude_code/providers/nvidia_nim/client.py
index 8afb4ed3..b8c77e79 100644
--- a/src/free_claude_code/providers/nvidia_nim/client.py
+++ b/src/free_claude_code/providers/nvidia_nim/client.py
@@ -8,7 +8,9 @@ from typing import Any
import openai
from loguru import logger
+from free_claude_code.application.model_metadata import ProviderModelInfo
from free_claude_code.config.nim import NimSettings
+from free_claude_code.config.provider_catalog import NVIDIA_NIM_DEFAULT_BASE
from free_claude_code.core.anthropic.models import MessagesRequest
from free_claude_code.core.failures import ExecutionFailure
from free_claude_code.core.reasoning import DEFAULT_REASONING_POLICY, ReasoningPolicy
@@ -25,6 +27,7 @@ from free_claude_code.providers.openai_chat import (
OpenAIChatProvider,
)
+from .models import HOSTED_MODEL_EXCLUSIONS
from .native_tool_stream import normalize_nim_native_tool_stream
from .request_options import (
NIM_REQUEST_POLICY,
@@ -168,6 +171,15 @@ class NvidiaNimProvider(OpenAIChatProvider):
admission=admission,
)
+ async def list_model_infos(self) -> frozenset[ProviderModelInfo]:
+ """Filter known unavailable and non-chat models from hosted discovery."""
+ infos = await super().list_model_infos()
+ if self._base_url != NVIDIA_NIM_DEFAULT_BASE:
+ return infos
+ return frozenset(
+ info for info in infos if info.model_id not in HOSTED_MODEL_EXCLUSIONS
+ )
+
def _nim_error_bodies(error: Exception) -> tuple[Mapping[str, Any], ...]:
body = getattr(error, "body", None)
diff --git a/src/free_claude_code/providers/nvidia_nim/models.py b/src/free_claude_code/providers/nvidia_nim/models.py
new file mode 100644
index 00000000..41accc4d
--- /dev/null
+++ b/src/free_claude_code/providers/nvidia_nim/models.py
@@ -0,0 +1,72 @@
+"""Maintained exclusions from NVIDIA's hosted chat model discovery.
+
+Recheck these exact IDs when updating the list. Observed 404s describe the
+tested hosted route and account, not permanent worldwide retirement. Remove
+an exclusion when hosted chat is verified to work again. New IDs remain visible.
+"""
+
+HOSTED_MODEL_EXCLUSIONS = frozenset(
+ {
+ # Model-specific 404 on the tested hosted routes, checked 2026-10-13.
+ "01-ai/yi-large",
+ "adept/fuyu-8b",
+ "ai21labs/jamba-1.5-large-instruct",
+ "aisingapore/sea-lion-7b-instruct",
+ "bigcode/starcoder2-15b",
+ "databricks/dbrx-instruct",
+ "deepseek-ai/deepseek-coder-6.7b-instruct",
+ "google/codegemma-1.1-7b",
+ "google/codegemma-7b",
+ "google/deplot",
+ "google/gemma-2b",
+ "google/gemma-3-12b-it",
+ "google/gemma-3-4b-it",
+ "google/recurrentgemma-2b",
+ "ibm/granite-3.0-3b-a800m-instruct",
+ "ibm/granite-3.0-8b-instruct",
+ "ibm/granite-34b-code-instruct",
+ "ibm/granite-8b-code-instruct",
+ "meta/codellama-70b",
+ "meta/llama2-70b",
+ "meta/muse-glimmer-30b",
+ "microsoft/kosmos-2",
+ "microsoft/phi-3-vision-128k-instruct",
+ "microsoft/phi-3.5-moe-instruct",
+ "mistralai/codestral-22b-instruct-v0.1",
+ "mistralai/mistral-7b-instruct-v0.3",
+ "mistralai/mistral-large",
+ "mistralai/mistral-large-2-instruct",
+ "mistralai/mixtral-8x22b-v0.1",
+ "moonshotai/kimi-k2.6",
+ "nv-mistralai/mistral-nemo-12b-instruct",
+ "nvidia/cosmos-reason2-8b",
+ "nvidia/embed-qa-4",
+ "nvidia/llama-3.1-nemotron-51b-instruct",
+ "nvidia/llama-3.1-nemotron-70b-instruct",
+ "nvidia/llama-3.1-nemotron-ultra-253b-v1",
+ "nvidia/llama-3.2-nemoretriever-1b-vlm-embed-v1",
+ "nvidia/llama-3.2-nv-embedqa-1b-v1",
+ "nvidia/llama3-chatqa-1.5-70b",
+ "nvidia/mistral-nemo-minitron-8b-8k-instruct",
+ "nvidia/nemotron-4-340b-instruct",
+ "nvidia/nemotron-4-340b-reward",
+ "nvidia/nemotron-nano-3-30b-a3b",
+ "nvidia/neva-22b",
+ "nvidia/nv-embedqa-mistral-7b-v2",
+ "nvidia/nvclip",
+ "nvidia/riva-translate-4b-instruct",
+ "nvidia/vila",
+ "snowflake/arctic-embed-l",
+ "writer/palmyra-creative-122b",
+ "writer/palmyra-fin-70b-32k",
+ "writer/palmyra-med-70b",
+ "writer/palmyra-med-70b-32k",
+ "zyphra/zamba2-7b-instruct",
+ # Explicit retirement 410, checked 2026-10-13.
+ "nvidia/ising-calibration-1-35b-a3b",
+ # Non-chat services, checked 2026-10-13.
+ "nvidia/ai-synthetic-video-detector",
+ "nvidia/llama-nemotron-embed-vl-1b-v2",
+ "nvidia/nemotron-3-embed-1b",
+ }
+)
diff --git a/tests/providers/test_model_discovery.py b/tests/providers/test_model_discovery.py
index a5adbd53..6fa30b37 100644
--- a/tests/providers/test_model_discovery.py
+++ b/tests/providers/test_model_discovery.py
@@ -110,6 +110,102 @@ async def test_nim_lists_openai_compatible_model_infos() -> None:
assert await provider.list_model_infos() == _infos("nvidia/model")
+ "provider_id,provider_base_url,filtered",
+ [
+ ("nvidia_nim", NVIDIA_NIM_DEFAULT_BASE, True),
+ ("nvidia_nim", NVIDIA_NIM_DEFAULT_BASE + "/", True),
+ ("nvidia_nim", "https://private-nim.example/v1", False),
+ ("deepseek", NVIDIA_NIM_DEFAULT_BASE, False),
+ ],
+)
+async def test_nim_exclusions_apply_only_to_hosted_nim_discovery(
+ provider_id, provider_base_url, filtered
+) -> None:
+ config = make_provider_config(api_key="test-key", base_url=provider_base_url)
+ with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
+ provider = (
+ NvidiaNimProvider(
+ config, nim_settings=NimSettings(), admission=immediate_admission()
+ )
+ if provider_id == "nvidia_nim"
+ else DeepSeekProvider(config, admission=immediate_admission())
+ )
+ excluded = (
+ "01-ai/yi-large", # Observed model-specific 404.
+ "meta/muse-glimmer-30b",
+ "nvidia/cosmos-reason2-8b",
+ "nvidia/ising-calibration-1-35b-a3b", # Explicit retirement 410.
+ "nvidia/llama-nemotron-embed-vl-1b-v2", # Working embedding service.
+ "nvidia/ai-synthetic-video-detector",
+ )
+ retained = (
+ "nvidia/nemotron-3-super-120b-a12b",
+ "nvidia/future-chat-model",
+ "01-ai/yi-large-successor", # Exact IDs, not substring matching.
+ "deepseek-ai/deepseek-v4.1-flash", # Timeouts are not exclusions.
+ "meta/llama-3.2-90b-vision-instruct",
+ "meta/llama-guard-4-12b",
+ "nvidia/llama-3.1-nemoguard-8b-content-safety",
+ "nvidia/llama-3.1-nemoguard-8b-topic-control", # Upstream 500.
+ )
+ ids = (*excluded, *retained)
+ with patch.object(
+ provider._client.models,
+ "list",
+ new_callable=AsyncMock,
+ return_value=SimpleNamespace(data=[SimpleNamespace(id=id_) for id_ in ids]),
+ ) as mock_list:
+ assert await provider.list_model_infos() == _infos(
+ *(retained if filtered else ids)
+ )
+ mock_list.assert_awaited_once_with()
+
+
+async def test_nim_exclusions_preserve_retained_metadata() -> None:
+ config = make_provider_config(api_key="test-key", base_url=NVIDIA_NIM_DEFAULT_BASE)
+ with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
+ provider = NvidiaNimProvider(
+ config, nim_settings=NimSettings(), admission=immediate_admission()
+ )
+ retained = ProviderModelInfo(
+ "nvidia/future-chat-model",
+ supports_thinking=True,
+ context_window_tokens=131072,
+ max_output_tokens=8192,
+ )
+ with patch.object(
+ OpenAIChatProvider,
+ "list_model_infos",
+ return_value=frozenset({ProviderModelInfo("01-ai/yi-large"), retained}),
+ ):
+ infos = await provider.list_model_infos()
+ assert len(infos) == 1
+ assert next(iter(infos)) is retained
+
+
+async def test_nim_exclusions_preserve_discovery_errors() -> None:
+ config = make_provider_config(api_key="test-key", base_url=NVIDIA_NIM_DEFAULT_BASE)
+ with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
+ provider = NvidiaNimProvider(
+ config, nim_settings=NimSettings(), admission=immediate_admission()
+ )
+ with (
+ patch.object(
+ provider._client.models,
+ "list",
+ new_callable=AsyncMock,
+ return_value=SimpleNamespace(data=[SimpleNamespace(id=None)]),
+ ) as mock_list,
+ pytest.raises(ModelListResponseError),
+ ):
+ await provider.list_model_infos()
+ mock_list.assert_awaited_once_with()
+
+
@pytest.mark.asyncio
@pytest.mark.parametrize(
"provider",