From 6a57936de328ec010dea19af2834533104f277c3 Mon Sep 17 00:00:00 2001 From: jhpiedrahitao Date: Mon, 5 Oct 2026 15:36:43 -0500 Subject: [PATCH] feat(sambanova): add MiniMax-M3, fix max output tokens, remove retired models --- ...odel_prices_and_context_window_backup.json | 41 +++++++------------ model_prices_and_context_window.json | 41 +++++++------------ tests/unit/test_sambanova_model_metadata.py | 39 ++++++++++++++++++ 3 files changed, 67 insertions(+), 54 deletions(-) diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 40fdf083cf8..972561251b0 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -45647,39 +45647,26 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "sambanova/DeepSeek-R1": { - "input_cost_per_token": 5e-06, + "sambanova/MiniMax-M3": { + "input_cost_per_token": 6e-07, "litellm_provider": "sambanova", - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_tokens": 32768, + "max_input_tokens": 1048576, + "max_output_tokens": 1048576, + "max_tokens": 1048576, "mode": "chat", - "output_cost_per_token": 7e-06, - "source": "https://cloud.sambanova.ai/plans/pricing" - }, - "sambanova/Llama-4-Maverick-17B-128E-Instruct": { - "input_cost_per_token": 6.3e-07, - "litellm_provider": "sambanova", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "metadata": { - "notes": "For vision models, images are converted to 6432 input tokens and are billed at that amount" - }, - "mode": "chat", - "output_cost_per_token": 1.8e-06, + "output_cost_per_token": 2.4e-06, "source": "https://cloud.sambanova.ai/plans/pricing", "supports_function_calling": true, + "supports_reasoning": true, "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "supports_tool_choice": true }, "sambanova/Meta-Llama-3.3-70B-Instruct": { "input_cost_per_token": 6e-07, "litellm_provider": "sambanova", "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, + "max_output_tokens": 3072, + "max_tokens": 3072, "mode": "chat", "output_cost_per_token": 1.2e-06, "source": "https://cloud.sambanova.ai/plans/pricing", @@ -45688,9 +45675,9 @@ "supports_tool_choice": true }, "sambanova/DeepSeek-V3.1": { - "max_tokens": 131072, + "max_tokens": 7168, "max_input_tokens": 131072, - "max_output_tokens": 131072, + "max_output_tokens": 7168, "input_cost_per_token": 3e-06, "output_cost_per_token": 4.5e-06, "litellm_provider": "sambanova", @@ -45714,9 +45701,9 @@ "source": "https://cloud.sambanova.ai/plans/pricing" }, "sambanova/DeepSeek-V3.2": { - "max_tokens": 32768, + "max_tokens": 7168, "max_input_tokens": 32768, - "max_output_tokens": 32768, + "max_output_tokens": 7168, "input_cost_per_token": 3e-06, "output_cost_per_token": 4.5e-06, "litellm_provider": "sambanova", diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 40fdf083cf8..972561251b0 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -45647,39 +45647,26 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "sambanova/DeepSeek-R1": { - "input_cost_per_token": 5e-06, + "sambanova/MiniMax-M3": { + "input_cost_per_token": 6e-07, "litellm_provider": "sambanova", - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_tokens": 32768, + "max_input_tokens": 1048576, + "max_output_tokens": 1048576, + "max_tokens": 1048576, "mode": "chat", - "output_cost_per_token": 7e-06, - "source": "https://cloud.sambanova.ai/plans/pricing" - }, - "sambanova/Llama-4-Maverick-17B-128E-Instruct": { - "input_cost_per_token": 6.3e-07, - "litellm_provider": "sambanova", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "metadata": { - "notes": "For vision models, images are converted to 6432 input tokens and are billed at that amount" - }, - "mode": "chat", - "output_cost_per_token": 1.8e-06, + "output_cost_per_token": 2.4e-06, "source": "https://cloud.sambanova.ai/plans/pricing", "supports_function_calling": true, + "supports_reasoning": true, "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "supports_tool_choice": true }, "sambanova/Meta-Llama-3.3-70B-Instruct": { "input_cost_per_token": 6e-07, "litellm_provider": "sambanova", "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, + "max_output_tokens": 3072, + "max_tokens": 3072, "mode": "chat", "output_cost_per_token": 1.2e-06, "source": "https://cloud.sambanova.ai/plans/pricing", @@ -45688,9 +45675,9 @@ "supports_tool_choice": true }, "sambanova/DeepSeek-V3.1": { - "max_tokens": 131072, + "max_tokens": 7168, "max_input_tokens": 131072, - "max_output_tokens": 131072, + "max_output_tokens": 7168, "input_cost_per_token": 3e-06, "output_cost_per_token": 4.5e-06, "litellm_provider": "sambanova", @@ -45714,9 +45701,9 @@ "source": "https://cloud.sambanova.ai/plans/pricing" }, "sambanova/DeepSeek-V3.2": { - "max_tokens": 32768, + "max_tokens": 7168, "max_input_tokens": 32768, - "max_output_tokens": 32768, + "max_output_tokens": 7168, "input_cost_per_token": 3e-06, "output_cost_per_token": 4.5e-06, "litellm_provider": "sambanova", diff --git a/tests/unit/test_sambanova_model_metadata.py b/tests/unit/test_sambanova_model_metadata.py index 20f34f9f3cc..17c7a9c591a 100644 --- a/tests/unit/test_sambanova_model_metadata.py +++ b/tests/unit/test_sambanova_model_metadata.py @@ -23,3 +23,42 @@ def test_sambanova_minimax_m27_model_info(): routed_model, provider, _, _ = get_llm_provider(model=model) assert routed_model == "MiniMax-M2.7" assert provider == "sambanova" + + +def _sambanova_models() -> dict: + json_path = Path(__file__).parents[2] / "model_prices_and_context_window.json" + with open(json_path) as f: + model_cost = json.load(f) + return {k: v for k, v in model_cost.items() if k.startswith("sambanova/")} + + +def test_sambanova_minimax_m3_model_info(): + info = _sambanova_models().get("sambanova/MiniMax-M3") + assert info is not None, "sambanova/MiniMax-M3 missing from model_prices_and_context_window.json" + assert info["litellm_provider"] == "sambanova" + assert info["mode"] == "chat" + assert info["max_input_tokens"] == 1048576 + assert info["input_cost_per_token"] == 6e-07 + assert info["output_cost_per_token"] == 2.4e-06 + assert info["supports_function_calling"] is True + assert info["supports_tool_choice"] is True + assert info["supports_reasoning"] is True + assert info["supports_response_schema"] is True + assert not info.get("supports_vision") + + routed_model, provider, _, _ = get_llm_provider(model="sambanova/MiniMax-M3") + assert routed_model == "MiniMax-M3" + assert provider == "sambanova" + + +def test_sambanova_max_output_tokens_match_live_catalog(): + models = _sambanova_models() + assert models["sambanova/DeepSeek-V3.1"]["max_output_tokens"] == 7168 + assert models["sambanova/DeepSeek-V3.2"]["max_output_tokens"] == 7168 + assert models["sambanova/Meta-Llama-3.3-70B-Instruct"]["max_output_tokens"] == 3072 + + +def test_sambanova_retired_models_removed(): + models = _sambanova_models() + assert "sambanova/DeepSeek-R1" not in models + assert "sambanova/Llama-4-Maverick-17B-128E-Instruct" not in models