diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index 90c2b44b741..52b1b5da939 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -74785,13 +74785,13 @@ }, "prism/deepseek-v4.1-flash": { "cache_read_input_token_cost": 7e-08, - "input_cost_per_token": 3e-07, + "input_cost_per_token": 1.7e-07, "litellm_provider": "prism", "max_input_tokens": 1000000, "max_output_tokens": 384000, "max_tokens": 384000, "mode": "chat", - "output_cost_per_token": 1.2e-06, + "output_cost_per_token": 6.3e-07, "source": "https://prisminference.com/pricing", "supported_endpoints": [ "/v1/chat/completions", @@ -74805,17 +74805,18 @@ "supports_reasoning": true, "supports_response_schema": true, "supports_system_messages": true, - "supports_tool_choice": true + "supports_tool_choice": true, + "supports_vision": true }, "prism/deepseek-v4-flash": { "cache_read_input_token_cost": 7e-08, - "input_cost_per_token": 1.4e-07, + "input_cost_per_token": 1.7e-07, "litellm_provider": "prism", "max_input_tokens": 1000000, - "max_output_tokens": 393216, - "max_tokens": 393216, + "max_output_tokens": 384000, + "max_tokens": 384000, "mode": "chat", - "output_cost_per_token": 2.8e-07, + "output_cost_per_token": 2.1e-07, "source": "https://prisminference.com/pricing", "supported_endpoints": [ "/v1/chat/completions", @@ -74829,6 +74830,7 @@ "supports_reasoning": true, "supports_response_schema": true, "supports_system_messages": true, - "supports_tool_choice": true + "supports_tool_choice": true, + "supports_vision": false } } diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 90c2b44b741..52b1b5da939 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -74785,13 +74785,13 @@ }, "prism/deepseek-v4.1-flash": { "cache_read_input_token_cost": 7e-08, - "input_cost_per_token": 3e-07, + "input_cost_per_token": 1.7e-07, "litellm_provider": "prism", "max_input_tokens": 1000000, "max_output_tokens": 384000, "max_tokens": 384000, "mode": "chat", - "output_cost_per_token": 1.2e-06, + "output_cost_per_token": 6.3e-07, "source": "https://prisminference.com/pricing", "supported_endpoints": [ "/v1/chat/completions", @@ -74805,17 +74805,18 @@ "supports_reasoning": true, "supports_response_schema": true, "supports_system_messages": true, - "supports_tool_choice": true + "supports_tool_choice": true, + "supports_vision": true }, "prism/deepseek-v4-flash": { "cache_read_input_token_cost": 7e-08, - "input_cost_per_token": 1.4e-07, + "input_cost_per_token": 1.7e-07, "litellm_provider": "prism", "max_input_tokens": 1000000, - "max_output_tokens": 393216, - "max_tokens": 393216, + "max_output_tokens": 384000, + "max_tokens": 384000, "mode": "chat", - "output_cost_per_token": 2.8e-07, + "output_cost_per_token": 2.1e-07, "source": "https://prisminference.com/pricing", "supported_endpoints": [ "/v1/chat/completions", @@ -74829,6 +74830,7 @@ "supports_reasoning": true, "supports_response_schema": true, "supports_system_messages": true, - "supports_tool_choice": true + "supports_tool_choice": true, + "supports_vision": false } } diff --git a/tests/test_litellm/llms/openai_like/test_prism_provider.py b/tests/test_litellm/llms/openai_like/test_prism_provider.py index f6b56612fd1..fbfa3abad37 100644 --- a/tests/test_litellm/llms/openai_like/test_prism_provider.py +++ b/tests/test_litellm/llms/openai_like/test_prism_provider.py @@ -52,17 +52,17 @@ def test_prism_provider_keeps_explicit_credentials(monkeypatch: pytest.MonkeyPat @pytest.mark.parametrize( - ("model", "input_cost", "output_cost", "max_output_tokens"), + ("model", "input_cost", "output_cost", "supports_vision"), [ - ("prism/deepseek-v4.1-flash", 0.30, 1.20, 384_000), - ("prism/deepseek-v4-flash", 0.14, 0.28, 393_216), + ("prism/deepseek-v4.1-flash", 0.17, 0.63, True), + ("prism/deepseek-v4-flash", 0.17, 0.21, False), ], ) def test_prism_model_cost_and_capabilities( model: str, input_cost: float, output_cost: float, - max_output_tokens: int, + supports_vision: bool, ): from litellm.cost_calculator import cost_per_token @@ -78,11 +78,12 @@ def test_prism_model_cost_and_capabilities( assert completion_cost == pytest.approx(output_cost) assert model_info["cache_read_input_token_cost"] == pytest.approx(7e-08) assert model_info["max_input_tokens"] == 1_000_000 - assert model_info["max_output_tokens"] == max_output_tokens + assert model_info["max_output_tokens"] == 384_000 assert model_info["supports_function_calling"] is True assert model_info["supports_native_streaming"] is True assert model_info["supports_reasoning"] is True assert model_info["supports_response_schema"] is True + assert litellm.supports_vision(model) is supports_vision def test_prism_is_available_in_add_model_form():