diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index cc418c9c428..46220e1c29f 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -16679,8 +16679,8 @@ "input_cost_per_token": 6e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 3e-06, "source": "https://fireworks.ai/pricing", @@ -16693,8 +16693,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -16709,8 +16709,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17053,8 +17053,8 @@ "input_cost_per_token": 6e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 3e-06, "source": "https://fireworks.ai/pricing", @@ -17067,8 +17067,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17083,8 +17083,8 @@ "input_cost_per_token": 2e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17099,8 +17099,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17115,8 +17115,8 @@ "input_cost_per_token": 1.9e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -42477,8 +42477,8 @@ "input_cost_per_token": 2e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -42493,8 +42493,8 @@ "input_cost_per_token": 1.9e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index c4628fecdb8..314a076dc20 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -16679,8 +16679,8 @@ "input_cost_per_token": 6e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 3e-06, "source": "https://fireworks.ai/pricing", @@ -16693,8 +16693,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -16709,8 +16709,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17053,8 +17053,8 @@ "input_cost_per_token": 6e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 3e-06, "source": "https://fireworks.ai/pricing", @@ -17067,8 +17067,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17083,8 +17083,8 @@ "input_cost_per_token": 2e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17099,8 +17099,8 @@ "input_cost_per_token": 9.5e-07, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 4e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -17115,8 +17115,8 @@ "input_cost_per_token": 1.9e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -42598,8 +42598,8 @@ "input_cost_per_token": 2e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", @@ -42614,8 +42614,8 @@ "input_cost_per_token": 1.9e-06, "litellm_provider": "fireworks_ai", "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, "mode": "chat", "output_cost_per_token": 8e-06, "source": "https://docs.fireworks.ai/serverless/pricing", diff --git a/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_kimi_model_metadata.py b/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_kimi_model_metadata.py new file mode 100644 index 00000000000..5641439aa54 --- /dev/null +++ b/tests/test_litellm/llms/fireworks_ai/test_fireworks_ai_kimi_model_metadata.py @@ -0,0 +1,76 @@ +""" +Regression test for Fireworks Kimi K2.5 / K2.6 / K2.7 context and output limits. + +Fireworks publishes a 262144-token context window for every Kimi K2.5, K2.6 and +K2.7 model, but caps generation well below that. A previous bulk edit had flattened +max_output_tokens/max_tokens to 262144 (equal to the context window), which let the +pre-call context-window check admit requests asking for a full 262144-token +completion that Fireworks then rejects. These assertions pin the corrected per-alias +limits so a future bulk edit can't silently flatten them again. +""" + +import json +from importlib.resources import files + +import pytest + +CONTEXT_WINDOW = 262144 +OUTPUT_LIMIT = 32768 + +KIMI_ALIASES = ( + "fireworks_ai/kimi-k2p5", + "fireworks_ai/kimi-k2p6", + "fireworks_ai/kimi-k2p6-fast", + "fireworks_ai/kimi-k2p7-code", + "fireworks_ai/kimi-k2p7-code-fast", + "fireworks_ai/accounts/fireworks/models/kimi-k2p5", + "fireworks_ai/accounts/fireworks/models/kimi-k2p6", + "fireworks_ai/accounts/fireworks/models/kimi-k2p7-code", + "fireworks_ai/accounts/fireworks/routers/kimi-k2p6-fast", + "fireworks_ai/accounts/fireworks/routers/kimi-k2p7-code-fast", +) + + +@pytest.fixture(scope="module") +def use_local_model_cost_map(): + monkeypatch = pytest.MonkeyPatch() + monkeypatch.setenv("LITELLM_LOCAL_MODEL_COST_MAP", "True") + + import litellm + from litellm.utils import _invalidate_model_cost_lowercase_map + + original_model_cost = litellm.model_cost + litellm.model_cost = json.loads( + files("litellm") + .joinpath("model_prices_and_context_window_backup.json") + .read_text(encoding="utf-8") + ) + litellm.get_model_info.cache_clear() + _invalidate_model_cost_lowercase_map() + try: + yield litellm + finally: + litellm.model_cost = original_model_cost + litellm.get_model_info.cache_clear() + _invalidate_model_cost_lowercase_map() + monkeypatch.undo() + + +@pytest.mark.parametrize("alias", KIMI_ALIASES) +def test_fireworks_kimi_raw_cost_entry_limits(use_local_model_cost_map, alias): + entry = use_local_model_cost_map.model_cost[alias] + + assert entry["litellm_provider"] == "fireworks_ai" + assert entry["max_input_tokens"] == CONTEXT_WINDOW + assert entry["max_output_tokens"] == OUTPUT_LIMIT + assert entry["max_tokens"] == OUTPUT_LIMIT + assert entry["max_output_tokens"] < entry["max_input_tokens"] + + +@pytest.mark.parametrize("alias", KIMI_ALIASES) +def test_fireworks_kimi_get_model_info_limits(use_local_model_cost_map, alias): + model_info = use_local_model_cost_map.get_model_info(model=alias) + + assert model_info["max_input_tokens"] == CONTEXT_WINDOW + assert model_info["max_output_tokens"] == OUTPUT_LIMIT + assert model_info["max_tokens"] == OUTPUT_LIMIT diff --git a/tests/test_litellm/test_utils.py b/tests/test_litellm/test_utils.py index b22e69f0942..3b2bc7e647d 100644 --- a/tests/test_litellm/test_utils.py +++ b/tests/test_litellm/test_utils.py @@ -4432,7 +4432,7 @@ _FIREWORKS_MODELS = [ 4e-06, 1.9e-07, 262144, - 262144, + 32768, True, True, ), @@ -4442,7 +4442,7 @@ _FIREWORKS_MODELS = [ 8e-06, 3.8e-07, 262144, - 262144, + 32768, True, True, ), @@ -4452,7 +4452,7 @@ _FIREWORKS_MODELS = [ 4e-06, 1.6e-07, 262144, - 262144, + 32768, True, True, ), @@ -4462,7 +4462,7 @@ _FIREWORKS_MODELS = [ 8e-06, 3e-07, 262144, - 262144, + 32768, True, True, ),