@pydantic/genai-prices 0.0.38 → 0.0.40

package/dist/index.js

@@ -463,13 +463,13 @@ const x = [
     ],
     models: [
       {
-        id: "meta.llama3-8b-instruct-v1%3A0",
+        id: "meta.llama3-8b-instruct-v1:0",
         match: {
-          equals: "meta.llama3-8b-instruct-v1%3A0"
+          contains: "llama3-8b-instruct-v1"
         },
         prices: {
-          input_mtok: 220,
-          output_mtok: 720
+          input_mtok: 0.3,
+          output_mtok: 0.6
         }
       },
       {
@@ -479,15 +479,13 @@ const x = [
         match: {
           or: [
             {
-              equals: "nova-lite-v1"
-            },
-            {
-              equals: "amazon.nova-lite-v1:0"
+              contains: "nova-lite-v1"
             }
           ]
         },
         prices: {
           input_mtok: 0.06,
+          cache_read_mtok: 0.015,
           output_mtok: 0.24
         }
       },
@@ -498,15 +496,13 @@ const x = [
         match: {
           or: [
             {
-              equals: "nova-micro-v1"
-            },
-            {
-              equals: "amazon.nova-micro-v1:0"
+              contains: "nova-micro-v1"
             }
           ]
         },
         prices: {
           input_mtok: 0.035,
+          cache_read_mtok: 875e-5,
           output_mtok: 0.14
         }
       },
@@ -517,15 +513,13 @@ const x = [
         match: {
           or: [
             {
-              equals: "nova-pro-v1"
-            },
-            {
-              equals: "amazon.nova-pro-v1:0"
+              contains: "nova-pro-v1"
             }
           ]
         },
         prices: {
           input_mtok: 0.8,
+          cache_read_mtok: 0.2,
           output_mtok: 3.2
         }
       }
@@ -1302,6 +1296,135 @@ const x = [
       }
     ]
   },
+  {
+    id: "cerebras",
+    name: "Cerebras",
+    pricing_urls: ["https://www.cerebras.ai/pricing#pricing", "https://inference-docs.cerebras.ai/models/openai-oss"],
+    api_pattern: "https://api\\.cerebras\\.ai",
+    model_match: {
+      contains: "cerebras"
+    },
+    provider_match: {
+      contains: "cerebras"
+    },
+    extractors: [
+      {
+        api_flavor: "chat",
+        root: "usage",
+        model_path: "model",
+        mappings: [
+          {
+            path: "prompt_tokens",
+            dest: "input_tokens",
+            required: !0
+          },
+          {
+            path: "completion_tokens",
+            dest: "output_tokens",
+            required: !0
+          }
+        ]
+      }
+    ],
+    models: [
+      {
+        id: "gpt-oss-120b",
+        name: "GPT-OSS 120B",
+        description: "OpenAI's flagship open source model, built on a Mixture-of-Experts (MoE) architecture with 120 billion parameters and 128 experts. Delivers frontier reasoning capabilities with record-breaking inference speeds on Cerebras hardware (~3,000 tokens/second).",
+        match: {
+          or: [
+            {
+              equals: "gpt-oss-120b"
+            },
+            {
+              starts_with: "cerebras/gpt-oss-120b"
+            },
+            {
+              starts_with: "cerebras:gpt-oss-120b"
+            }
+          ]
+        },
+        context_window: 131072,
+        price_comments: "Developer tier pricing. Free tier: 65k context, Paid tier: 131k context.",
+        prices: {
+          input_mtok: 0.35,
+          output_mtok: 0.75
+        }
+      },
+      {
+        id: "llama-3.3-70b",
+        name: "Llama 3.3 70B",
+        description: "Meta's enhanced 70B model delivering 405B-level accuracy. Optimized for chat, coding, instruction following, mathematics, and reasoning with high-speed inference on Cerebras hardware (~2,100 tokens/second).",
+        match: {
+          or: [
+            {
+              equals: "llama-3.3-70b"
+            },
+            {
+              starts_with: "cerebras/llama-3.3-70b"
+            },
+            {
+              starts_with: "cerebras:llama-3.3-70b"
+            }
+          ]
+        },
+        context_window: 128e3,
+        price_comments: "Developer tier pricing. Free tier: 65k context, Paid tier: 128k context.",
+        prices: {
+          input_mtok: 0.85,
+          output_mtok: 1.2
+        }
+      },
+      {
+        id: "llama3.1-8b",
+        name: "Llama 3.1 8B",
+        description: "Meta's Llama 3.1 8B model for general-purpose tasks including chat, coding, and instruction following. Optimized for fast inference on Cerebras hardware (~2,200 tokens/second).",
+        match: {
+          or: [
+            {
+              equals: "llama3.1-8b"
+            },
+            {
+              starts_with: "cerebras/llama3.1-8b"
+            },
+            {
+              starts_with: "cerebras:llama3.1-8b"
+            }
+          ]
+        },
+        context_window: 32768,
+        price_comments: "Developer tier pricing. Free tier: 8k context, Paid tier: 32k context.",
+        prices: {
+          input_mtok: 0.1,
+          output_mtok: 0.1
+        }
+      },
+      {
+        id: "qwen-3-32b",
+        name: "Qwen 3 32B",
+        description: "Qwen's 32B parameter model with enhanced reasoning and coding capabilities. Supports both standard and reasoning modes for complex tasks, with fast inference speeds on Cerebras hardware (~2,600 tokens/second).",
+        match: {
+          or: [
+            {
+              equals: "qwen-3-32b"
+            },
+            {
+              starts_with: "cerebras/qwen-3-32b"
+            },
+            {
+              starts_with: "cerebras:qwen-3-32b"
+            }
+          ]
+        },
+        context_window: 131072,
+        price_comments: "Developer tier pricing. Free tier: 65k context, Paid tier: 131k context.",
+        prices: {
+          input_mtok: 0.4,
+          output_mtok: 0.8
+        }
+      }
+    ]
+  },
   {
     id: "cohere",
     name: "Cohere",
@@ -1699,7 +1822,12 @@ Compared to other leading proprietary and open-weights models Command A delivers
           {
             path: "promptTokenCount",
             dest: "input_tokens",
-            required: !0
+            required: !1
+          },
+          {
+            path: "cachedContentTokenCount",
+            dest: "cache_read_tokens",
+            required: !1
           },
           {
             path: [
@@ -1708,17 +1836,17 @@ Compared to other leading proprietary and open-weights models Command A delivers
                 type: "array-match",
                 field: "modality",
                 match: {
-                  equals: "TEXT"
+                  equals: "AUDIO"
                 }
               },
               "tokenCount"
             ],
-            dest: "cache_read_tokens",
+            dest: "cache_audio_read_tokens",
             required: !1
           },
           {
             path: [
-              "cacheTokensDetails",
+              "promptTokensDetails",
               {
                 type: "array-match",
                 field: "modality",
@@ -1728,12 +1856,12 @@ Compared to other leading proprietary and open-weights models Command A delivers
               },
               "tokenCount"
             ],
-            dest: "cache_audio_read_tokens",
+            dest: "input_audio_tokens",
             required: !1
           },
           {
             path: [
-              "promptTokensDetails",
+              "candidatesTokensDetails",
               {
                 type: "array-match",
                 field: "modality",
@@ -1743,18 +1871,23 @@ Compared to other leading proprietary and open-weights models Command A delivers
               },
               "tokenCount"
             ],
-            dest: "input_audio_tokens",
+            dest: "output_audio_tokens",
             required: !1
           },
           {
             path: "candidatesTokenCount",
             dest: "output_tokens",
-            required: !0
+            required: !1
           },
           {
             path: "thoughtsTokenCount",
             dest: "output_tokens",
             required: !1
+          },
+          {
+            path: "toolUsePromptTokenCount",
+            dest: "output_tokens",
+            required: !1
           }
         ]
       },
@@ -2400,6 +2533,16 @@ Compared to other leading proprietary and open-weights models Command A delivers
           output_mtok: 0.99
         }
       },
+      {
+        id: "gemma-7b-it",
+        match: {
+          equals: "gemma-7b-it"
+        },
+        prices: {
+          input_mtok: 0.07,
+          output_mtok: 0.07
+        }
+      },
       {
         id: "gemma2-9b-it",
         name: "Gemma 2 9B 8k",
@@ -2418,6 +2561,26 @@ Compared to other leading proprietary and open-weights models Command A delivers
           output_mtok: 0.2
         }
       },
+      {
+        id: "llama-3.1-405b-reasoning",
+        match: {
+          equals: "llama-3.1-405b-reasoning"
+        },
+        prices: {
+          input_mtok: 0.59,
+          output_mtok: 0.79
+        }
+      },
+      {
+        id: "llama-3.1-70b-versatile",
+        match: {
+          equals: "llama-3.1-70b-versatile"
+        },
+        prices: {
+          input_mtok: 0.59,
+          output_mtok: 0.79
+        }
+      },
       {
         id: "llama-3.1-8b-instant",
         name: "Llama 3.1 8B Instant 128k",
@@ -2429,6 +2592,76 @@ Compared to other leading proprietary and open-weights models Command A delivers
           output_mtok: 0.08
         }
       },
+      {
+        id: "llama-3.2-11b-text-preview",
+        match: {
+          equals: "llama-3.2-11b-text-preview"
+        },
+        prices: {
+          input_mtok: 0.18,
+          output_mtok: 0.18
+        }
+      },
+      {
+        id: "llama-3.2-11b-vision-preview",
+        match: {
+          equals: "llama-3.2-11b-vision-preview"
+        },
+        prices: {
+          input_mtok: 0.18,
+          output_mtok: 0.18
+        }
+      },
+      {
+        id: "llama-3.2-1b-preview",
+        match: {
+          equals: "llama-3.2-1b-preview"
+        },
+        prices: {
+          input_mtok: 0.04,
+          output_mtok: 0.04
+        }
+      },
+      {
+        id: "llama-3.2-3b-preview",
+        match: {
+          equals: "llama-3.2-3b-preview"
+        },
+        prices: {
+          input_mtok: 0.06,
+          output_mtok: 0.06
+        }
+      },
+      {
+        id: "llama-3.2-90b-text-preview",
+        match: {
+          equals: "llama-3.2-90b-text-preview"
+        },
+        prices: {
+          input_mtok: 0.9,
+          output_mtok: 0.9
+        }
+      },
+      {
+        id: "llama-3.2-90b-vision-preview",
+        match: {
+          equals: "llama-3.2-90b-vision-preview"
+        },
+        prices: {
+          input_mtok: 0.9,
+          output_mtok: 0.9
+        }
+      },
+      {
+        id: "llama-3.3-70b-specdec",
+        match: {
+          equals: "llama-3.3-70b-specdec"
+        },
+        prices: {
+          input_mtok: 0.59,
+          output_mtok: 0.99
+        }
+      },
       {
         id: "llama-3.3-70b-versatile",
         name: "Llama 3.3 70B Versatile 128k",
@@ -2440,6 +2673,16 @@ Compared to other leading proprietary and open-weights models Command A delivers
           output_mtok: 0.79
         }
       },
+      {
+        id: "llama-guard-3-8b",
+        match: {
+          equals: "llama-guard-3-8b"
+        },
+        prices: {
+          input_mtok: 0.2,
+          output_mtok: 0.2
+        }
+      },
       {
         id: "llama2-70b-4096",
         match: {
@@ -2525,15 +2768,43 @@ Compared to other leading proprietary and open-weights models Command A delivers
           output_mtok: 0.2
         }
       },
+      {
+        id: "mistral-saba-24b",
+        match: {
+          equals: "mistral-saba-24b"
+        },
+        prices: {
+          input_mtok: 0.79,
+          output_mtok: 0.79
+        }
+      },
+      {
+        id: "mixtral-8x7b-32768",
+        match: {
+          equals: "mixtral-8x7b-32768"
+        },
+        prices: {
+          input_mtok: 0.24,
+          output_mtok: 0.24
+        }
+      },
       {
         id: "moonshotai/kimi-k2-instruct",
         name: "Kimi K2 1T 128k",
         match: {
-          equals: "moonshotai/kimi-k2-instruct"
+          or: [
+            {
+              equals: "moonshotai/kimi-k2-instruct"
+            },
+            {
+              equals: "moonshotai/kimi-k2-instruct-0905"
+            }
+          ]
         },
         context_window: 131072,
         prices: {
           input_mtok: 1,
+          cache_read_mtok: 0.5,
           output_mtok: 3
         }
       },
@@ -2543,12 +2814,20 @@ Compared to other leading proprietary and open-weights models Command A delivers
 120 billion parameters and 128 experts.
 `,
         match: {
-          equals: "openai/gpt-oss-120b"
+          or: [
+            {
+              equals: "openai/gpt-oss-120b"
+            },
+            {
+              equals: "openai/gpt-oss-safeguard-20b"
+            }
+          ]
         },
         context_window: 131072,
         prices: {
           input_mtok: 0.15,
-          output_mtok: 0.75
+          cache_read_mtok: 0.075,
+          output_mtok: 0.6
         }
       },
       {
@@ -2561,8 +2840,9 @@ Compared to other leading proprietary and open-weights models Command A delivers
         },
         context_window: 131072,
         prices: {
-          input_mtok: 0.1,
-          output_mtok: 0.5
+          input_mtok: 0.075,
+          cache_read_mtok: 0.0375,
+          output_mtok: 0.3
         }
       },
       {
@@ -3376,6 +3656,18 @@ Compared to other leading proprietary and open-weights models Command A delivers
             required: !0
           }
         ]
+      },
+      {
+        api_flavor: "embeddings",
+        root: "usage",
+        model_path: "model",
+        mappings: [
+          {
+            path: "prompt_tokens",
+            dest: "input_tokens",
+            required: !0
+          }
+        ]
       }
     ],
     models: [
@@ -3438,6 +3730,17 @@ Compared to other leading proprietary and open-weights models Command A delivers
           output_mtok: 6
         }
       },
+      {
+        id: "computer-use",
+        name: "Computer use",
+        match: {
+          starts_with: "computer-use"
+        },
+        prices: {
+          input_mtok: 3,
+          output_mtok: 12
+        }
+      },
       {
         id: "curie",
         match: {
@@ -4087,11 +4390,59 @@ Compared to other leading proprietary and open-weights models Command A delivers
             }
           ]
         },
+        context_window: 4e5,
         prices: {
           input_mtok: 15,
           output_mtok: 120
         }
       },
+      {
+        id: "gpt-5.1",
+        name: "GPT-5.1",
+        description: "The best model for coding and agentic tasks across industries",
+        match: {
+          or: [
+            {
+              equals: "gpt-5.1"
+            },
+            {
+              equals: "gpt-5.1-2025-11-13"
+            },
+            {
+              equals: "gpt-5.1-codex"
+            },
+            {
+              equals: "gpt-5.1-chat-latest"
+            }
+          ]
+        },
+        context_window: 4e5,
+        prices: {
+          input_mtok: 1.25,
+          cache_read_mtok: 0.125,
+          output_mtok: 10
+        }
+      },
+      {
+        id: "gpt-5.1-codex-mini",
+        name: "GPT-5.1 Codex Mini",
+        match: {
+          or: [
+            {
+              equals: "gpt-5.1-codex-mini"
+            },
+            {
+              equals: "gpt-5.1-mini"
+            }
+          ]
+        },
+        context_window: 4e5,
+        prices: {
+          input_mtok: 0.25,
+          cache_read_mtok: 0.025,
+          output_mtok: 2
+        }
+      },
       {
         id: "gpt-realtime",
         match: {
@@ -11600,7 +11951,7 @@ function L(t, e) {
   const a = e.toLowerCase().trim(), i = t.find((o) => o.id === a);
   return i || t.find((o) => o.provider_match && h(o.provider_match, a));
 }
-function y(t, { modelId: e, providerApiUrl: a, providerId: i }) {
+function B(t, { modelId: e, providerApiUrl: a, providerId: i }) {
   if (i)
     return L(t, i);
   if (a)
@@ -11618,22 +11969,22 @@ function S(t) {
     e !== null && (g = e);
   })) : (b = Promise.resolve(t), g = t));
 }
-function A(t) {
+function C(t) {
   f = t;
 }
-function N(t) {
+function Q(t) {
   t({
-    onCalc: A,
+    onCalc: C,
     remoteDataUrl: G,
     setProviderData: S
   });
 }
-function Q() {
+function N() {
   return b;
 }
 function O(t, e, a) {
   f?.();
-  const i = e.toLowerCase().trim(), o = a?.provider ?? y(g, { modelId: i, providerApiUrl: a?.providerApiUrl, providerId: a?.providerId });
+  const i = e.toLowerCase().trim(), o = a?.provider ?? B(g, { modelId: i, providerApiUrl: a?.providerApiUrl, providerId: a?.providerId });
   if (!o) return null;
   const m = T(o.models, i);
   if (!m) return null;
@@ -11646,10 +11997,10 @@ function O(t, e, a) {
     ...n
   };
 }
-function E(t) {
-  return f?.(), y(g, t);
+function D(t) {
+  return f?.(), B(g, t);
 }
-function D(t, e, a) {
+function E(t, e, a) {
   if (a = a ?? "default", !t.extractors)
     throw new Error("No extraction logic defined for this provider");
   const i = t.extractors.find((n) => n.api_flavor === a);
@@ -11659,7 +12010,7 @@ function D(t, e, a) {
   }
   if (!_.guard(e))
     throw new Error(`Expected response data to be a mapping object, got ${c(e)}`);
-  const o = q(i.model_path, e, C, !1, []), m = B(i.root), r = q(m, e, _, !0, []), s = {};
+  const o = q(i.model_path, e, R, !1, []), m = y(i.root), r = q(m, e, _, !0, []), s = {};
   for (const n of i.mappings) {
     const u = q(n.path, r, z, n.required, m);
     if (u !== null) {
@@ -11672,7 +12023,7 @@ function D(t, e, a) {
   return { model: o, usage: s };
 }
 function q(t, e, a, i, o) {
-  const [m, ...r] = B(t).reverse();
+  const [m, ...r] = y(t).reverse();
   if (typeof m != "string")
     throw new Error(`Expected last step of path to be a string, got ${c(m)}`);
   r.reverse();
@@ -11681,7 +12032,7 @@ function q(t, e, a, i, o) {
   for (const p of r) {
     if (n.push(p), typeof p == "object")
       if (Array.isArray(s))
-        s = R(p, s);
+        s = A(p, s);
       else {
         if (i)
           throw new Error(`Expected \`${d(o, n)}\` value to be a mapping, got ${c(s)}`);
@@ -11718,7 +12069,7 @@ function q(t, e, a, i, o) {
     throw n.push(m), new Error(`Expected \`${d(o, n)}\` value to be a ${a.name}, got ${c(u)}`);
   return null;
 }
-function R(t, e) {
+function A(t, e) {
   for (const a of e)
     if (_.guard(a)) {
       const i = a[t.field];
@@ -11726,7 +12077,7 @@ function R(t, e) {
         return a;
     }
 }
-function B(t) {
+function y(t) {
   return Array.isArray(t) ? [...t] : [t];
 }
 function c(t) {
@@ -11735,7 +12086,7 @@ function c(t) {
 const _ = {
   guard: (t) => c(t) === "mapping",
   name: "mapping"
-}, C = {
+}, R = {
   guard: (t) => typeof t == "string",
   name: "string"
 }, z = {
@@ -11745,8 +12096,8 @@ const _ = {
 export {
   G as REMOTE_DATA_JSON_URL,
   O as calcPrice,
-  D as extractUsage,
-  E as findProvider,
-  N as updatePrices,
-  Q as waitForUpdate
+  E as extractUsage,
+  D as findProvider,
+  Q as updatePrices,
+  N as waitForUpdate
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pydantic/genai-prices",
-  "version": "0.0.38",
+  "version": "0.0.40",
   "description": "Calculate prices for calling LLM inference APIs",
   "author": "Pydantic Team",
   "type": "module",