sync github copilot limits

add OpenRouter models for gpt-5.4 mini and gpt-5.4 nano

README.md

@@ -199,6 +199,17 @@ $ bun run dev
 
 And it'll open the frontend at http://localhost:3000
 
+### Manual testing with opencode
+
+You can manually check provider changes with opencode by:
+
+```bash
+$ bun install
+$ cd packages/web
+$ bun run build
+$ OPENCODE_MODELS_PATH="dist/_api.json" opencode
+```
+
 ### Questions?
 
 Open an issue if you need help or have questions about contributing.

package.json

@@ -18,7 +18,8 @@
     "validate": "bun ./packages/core/script/validate.ts",
     "helicone:generate": "bun ./packages/core/script/generate-helicone.ts",
     "venice:generate": "bun ./packages/core/script/generate-venice.ts",
-    "vercel:generate": "bun ./packages/core/script/generate-vercel.ts"
+    "vercel:generate": "bun ./packages/core/script/generate-vercel.ts",
+    "wandb:generate": "bun ./packages/core/script/generate-wandb.ts"
   },
   "dependencies": {
     "@cloudflare/workers-types": "^4.20250801.0",

packages/core/script/generate-vercel.ts

@@ -24,6 +24,7 @@ enum ModelType {
 
 enum SkipZeroFields {
   LimitContext = "limit.context",
+  LimitInput = "limit.input",
   LimitOutput = "limit.output",
 }
 
@@ -82,6 +83,7 @@ interface ExistingModel {
   };
   limit?: {
     context?: number;
+    input?: number;
     output?: number;
   };
   modalities?: {
@@ -112,6 +114,7 @@ interface MergedModel {
   };
   limit: {
     context: number;
+    input?: number;
     output: number;
   };
   modalities: {
@@ -232,6 +235,10 @@ async function loadExistingModel(filePath: string): Promise<ExistingModel | null
   }
 }
 
+function isOpenAIModel(modelId: string): boolean {
+  return modelId.startsWith("openai/");
+}
+
 function mergeModel(
   apiModel: z.infer<typeof VercelModel>,
   existing: ExistingModel | null,
@@ -281,6 +288,7 @@ function mergeModel(
     ...(status && { status }),
     limit: {
       context: contextLimit,
+      ...(isOpenAIModel(apiModel.id) && contextLimit > outputLimit && { input: contextLimit - outputLimit }),
       output: outputLimit,
     },
     modalities: {
@@ -362,6 +370,9 @@ function formatToml(model: MergedModel): string {
   lines.push("");
   lines.push(`[limit]`);
   lines.push(`context = ${formatNumber(model.limit.context)}`);
+  if (model.limit.input !== undefined) {
+    lines.push(`input = ${formatNumber(model.limit.input)}`);
+  }
   lines.push(`output = ${formatNumber(model.limit.output)}`);
 
   lines.push("");
@@ -435,6 +446,7 @@ function detectChanges(
   compare("cost.cache_read", existing.cost?.cache_read, merged.cost?.cache_read);
   compare("cost.cache_write", existing.cost?.cache_write, merged.cost?.cache_write);
   compare("limit.context", existing.limit?.context, merged.limit.context);
+  compare("limit.input", existing.limit?.input, merged.limit.input);
   compare("limit.output", existing.limit?.output, merged.limit.output);
   compare("modalities.input", existing.modalities?.input, merged.modalities.input);
 

packages/core/script/generate-wandb.ts

@@ -0,0 +1,525 @@
+#!/usr/bin/env bun
+
+import path from "node:path";
+import { mkdir } from "node:fs/promises";
+import { z } from "zod";
+import { ModelFamilyValues } from "../src/family.js";
+
+const API_ENDPOINT = "https://trace.wandb.ai/inference/analysis/artificialanalysis/models";
+
+const Pricing = z
+  .object({
+    prompt: z.string().optional(),
+    completion: z.string().optional(),
+    image: z.string().optional(),
+    request: z.string().optional(),
+    input_cache_reads: z.string().optional(),
+    input_cache_writes: z.string().optional(),
+  })
+  .passthrough();
+
+const WandbModel = z
+  .object({
+    id: z.string(),
+    name: z.string(),
+    created: z.number(),
+    input_modalities: z.array(z.string()),
+    output_modalities: z.array(z.string()),
+    context_length: z.number(),
+    max_output_length: z.number(),
+    pricing: Pricing.optional(),
+    supported_sampling_parameters: z.array(z.string()).default([]),
+    supported_features: z.array(z.string()).default([]),
+  })
+  .passthrough();
+
+const WandbResponse = z
+  .object({
+    data: z.array(WandbModel),
+  })
+  .strict();
+
+interface ExistingModel {
+  name?: string;
+  family?: string;
+  attachment?: boolean;
+  reasoning?: boolean;
+  tool_call?: boolean;
+  structured_output?: boolean;
+  temperature?: boolean;
+  knowledge?: string;
+  release_date?: string;
+  last_updated?: string;
+  open_weights?: boolean;
+  interleaved?: boolean | { field: string };
+  status?: string;
+  cost?: {
+    input?: number;
+    output?: number;
+    cache_read?: number;
+    cache_write?: number;
+  };
+  limit?: {
+    context?: number;
+    input?: number;
+    output?: number;
+  };
+  modalities?: {
+    input?: string[];
+    output?: string[];
+  };
+}
+
+interface MergedModel {
+  name: string;
+  family?: string;
+  attachment: boolean;
+  reasoning: boolean;
+  tool_call: boolean;
+  structured_output?: boolean;
+  temperature: boolean;
+  knowledge?: string;
+  release_date: string;
+  last_updated: string;
+  open_weights: boolean;
+  interleaved?: boolean | { field: string };
+  status?: string;
+  cost?: {
+    input: number;
+    output: number;
+    cache_read?: number;
+    cache_write?: number;
+  };
+  limit: {
+    context: number;
+    output: number;
+  };
+  modalities: {
+    input: Array<"text" | "audio" | "image" | "video" | "pdf">;
+    output: Array<"text" | "audio" | "image" | "video" | "pdf">;
+  };
+}
+
+interface Changes {
+  field: string;
+  oldValue: string;
+  newValue: string;
+}
+
+type SupportedModality = "text" | "audio" | "image" | "video" | "pdf";
+
+const modalityMap: Record<string, SupportedModality | undefined> = {
+  text: "text",
+  image: "image",
+  audio: "audio",
+  video: "video",
+  pdf: "pdf",
+  file: "pdf",
+  files: "pdf",
+};
+
+const openWeightsPrefixes = new Set([
+  "deepseek-ai/",
+  "meta-llama/",
+  "microsoft/",
+  "MiniMaxAI/",
+  "moonshotai/",
+  "nvidia/",
+  "OpenPipe/",
+  "Qwen/",
+  "zai-org/",
+]);
+
+function timestampToDate(timestamp: number): string {
+  return new Date(timestamp * 1000).toISOString().slice(0, 10);
+}
+
+function getTodayDate(): string {
+  return new Date().toISOString().slice(0, 10);
+}
+
+function formatNumber(n: number): string {
+  if (n >= 1000) {
+    return n.toString().replace(/\B(?=(\d{3})+(?!\d))/g, "_");
+  }
+  return n.toString();
+}
+
+function formatDecimal(n: number): string {
+  return Number(n.toFixed(6)).toString();
+}
+
+function priceToPerMillion(value: string): number {
+  return Number((parseFloat(value) * 1_000_000).toFixed(6));
+}
+
+function isSubstring(target: string, family: string): boolean {
+  return target.toLowerCase().includes(family.toLowerCase());
+}
+
+function matchesFamily(target: string, family: string): boolean {
+  const targetLower = target.toLowerCase();
+  const familyLower = family.toLowerCase();
+  let familyIdx = 0;
+
+  for (let i = 0; i < targetLower.length && familyIdx < familyLower.length; i++) {
+    if (targetLower[i] === familyLower[familyIdx]) {
+      familyIdx++;
+    }
+  }
+
+  return familyIdx === familyLower.length;
+}
+
+function inferFamily(modelId: string, modelName: string): string | undefined {
+  const sortedFamilies = [...ModelFamilyValues].sort((a, b) => b.length - a.length);
+
+  for (const family of sortedFamilies) {
+    if (isSubstring(modelId, family) || isSubstring(modelName, family)) {
+      return family;
+    }
+  }
+
+  for (const family of sortedFamilies) {
+    if (matchesFamily(modelId, family) || matchesFamily(modelName, family)) {
+      return family;
+    }
+  }
+
+  return undefined;
+}
+
+function normalizeName(apiModel: z.infer<typeof WandbModel>): string {
+  const stripped = apiModel.name.replace(/^[^:]+:\s*/, "").trim();
+  return stripped || path.basename(apiModel.id);
+}
+
+function inferReasoning(apiModel: z.infer<typeof WandbModel>): boolean {
+  const text = `${apiModel.id} ${apiModel.name}`.toLowerCase();
+  return text.includes("thinking") || /\br1\b/.test(text) || text.includes("reasoning");
+}
+
+function inferOpenWeights(modelId: string): boolean {
+  for (const prefix of openWeightsPrefixes) {
+    if (modelId.startsWith(prefix)) {
+      return true;
+    }
+  }
+
+  return false;
+}
+
+function normalizeModalities(values: string[]): SupportedModality[] {
+  const normalized = values
+    .map((value) => modalityMap[value.toLowerCase()])
+    .filter((value): value is SupportedModality => value !== undefined);
+
+  return [...new Set(normalized)];
+}
+
+async function loadExistingModel(filePath: string): Promise<ExistingModel | null> {
+  try {
+    const file = Bun.file(filePath);
+    if (!(await file.exists())) {
+      return null;
+    }
+
+    const toml = await import(filePath, { with: { type: "toml" } }).then((mod) => mod.default);
+    return toml as ExistingModel;
+  } catch (cause) {
+    console.warn(`Warning: Failed to parse existing file ${filePath}:`, cause);
+    return null;
+  }
+}
+
+function mergeModel(
+  apiModel: z.infer<typeof WandbModel>,
+  existing: ExistingModel | null,
+): MergedModel {
+  const featureSet = new Set(apiModel.supported_features);
+  const samplingSet = new Set(apiModel.supported_sampling_parameters);
+  const inputModalities = normalizeModalities(apiModel.input_modalities);
+  const outputModalities = normalizeModalities(apiModel.output_modalities);
+
+  const merged: MergedModel = {
+    name: existing?.name ?? normalizeName(apiModel),
+    family: existing?.family ?? inferFamily(apiModel.id, apiModel.name),
+    attachment: existing?.attachment ?? inputModalities.some((m) => m !== "text"),
+    reasoning: existing?.reasoning ?? inferReasoning(apiModel),
+    tool_call: existing?.tool_call ?? featureSet.has("tools"),
+    temperature: existing?.temperature ?? samplingSet.has("temperature"),
+    release_date: existing?.release_date ?? timestampToDate(apiModel.created),
+    last_updated: getTodayDate(),
+    open_weights: existing?.open_weights ?? inferOpenWeights(apiModel.id),
+    ...(existing?.structured_output !== undefined
+      ? { structured_output: existing.structured_output }
+      : featureSet.has("structured_outputs")
+        ? { structured_output: true }
+        : {}),
+    ...(existing?.knowledge ? { knowledge: existing.knowledge } : {}),
+    ...(existing?.interleaved !== undefined ? { interleaved: existing.interleaved } : {}),
+    ...(existing?.status ? { status: existing.status } : {}),
+    limit: {
+      context: apiModel.context_length > 0 ? apiModel.context_length : (existing?.limit?.context ?? 0),
+      output: apiModel.max_output_length > 0
+        ? apiModel.max_output_length
+        : (existing?.limit?.output ?? 0),
+    },
+    modalities: {
+      input: inputModalities.length > 0
+        ? inputModalities
+        : ((existing?.modalities?.input as SupportedModality[] | undefined) ?? ["text"]),
+      output: outputModalities.length > 0
+        ? outputModalities
+        : ((existing?.modalities?.output as SupportedModality[] | undefined) ?? ["text"]),
+    },
+  };
+
+  const prompt = apiModel.pricing?.prompt;
+  const completion = apiModel.pricing?.completion;
+  const cacheRead = apiModel.pricing?.input_cache_reads;
+  const cacheWrite = apiModel.pricing?.input_cache_writes;
+
+  if (prompt && completion) {
+    merged.cost = {
+      input: priceToPerMillion(prompt),
+      output: priceToPerMillion(completion),
+      ...(cacheRead && parseFloat(cacheRead) > 0
+        ? { cache_read: priceToPerMillion(cacheRead) }
+        : {}),
+      ...(cacheWrite && parseFloat(cacheWrite) > 0
+        ? { cache_write: priceToPerMillion(cacheWrite) }
+        : {}),
+    };
+  } else if (existing?.cost?.input !== undefined && existing.cost.output !== undefined) {
+    merged.cost = {
+      input: existing.cost.input,
+      output: existing.cost.output,
+      ...(existing.cost.cache_read !== undefined ? { cache_read: existing.cost.cache_read } : {}),
+      ...(existing.cost.cache_write !== undefined ? { cache_write: existing.cost.cache_write } : {}),
+    };
+  }
+
+  return merged;
+}
+
+function formatToml(model: MergedModel): string {
+  const lines: string[] = [];
+
+  lines.push(`name = "${model.name.replace(/"/g, '\\"')}"`);
+  if (model.family) {
+    lines.push(`family = "${model.family}"`);
+  }
+  lines.push(`release_date = "${model.release_date}"`);
+  lines.push(`last_updated = "${model.last_updated}"`);
+  lines.push(`attachment = ${model.attachment}`);
+  lines.push(`reasoning = ${model.reasoning}`);
+  if (model.structured_output !== undefined) {
+    lines.push(`structured_output = ${model.structured_output}`);
+  }
+  lines.push(`temperature = ${model.temperature}`);
+  lines.push(`tool_call = ${model.tool_call}`);
+  if (model.knowledge) {
+    lines.push(`knowledge = "${model.knowledge}"`);
+  }
+  lines.push(`open_weights = ${model.open_weights}`);
+  if (model.status) {
+    lines.push(`status = "${model.status}"`);
+  }
+
+  if (model.interleaved !== undefined) {
+    lines.push("");
+    if (model.interleaved === true) {
+      lines.push("interleaved = true");
+    } else {
+      lines.push("[interleaved]");
+      lines.push(`field = "${model.interleaved.field}"`);
+    }
+  }
+
+  if (model.cost) {
+    lines.push("");
+    lines.push("[cost]");
+    lines.push(`input = ${formatDecimal(model.cost.input)}`);
+    lines.push(`output = ${formatDecimal(model.cost.output)}`);
+    if (model.cost.cache_read !== undefined) {
+      lines.push(`cache_read = ${formatDecimal(model.cost.cache_read)}`);
+    }
+    if (model.cost.cache_write !== undefined) {
+      lines.push(`cache_write = ${formatDecimal(model.cost.cache_write)}`);
+    }
+  }
+
+  lines.push("");
+  lines.push("[limit]");
+  lines.push(`context = ${formatNumber(model.limit.context)}`);
+  lines.push(`output = ${formatNumber(model.limit.output)}`);
+
+  lines.push("");
+  lines.push("[modalities]");
+  lines.push(`input = [${model.modalities.input.map((m) => `"${m}"`).join(", ")}]`);
+  lines.push(`output = [${model.modalities.output.map((m) => `"${m}"`).join(", ")}]`);
+
+  return `${lines.join("\n")}\n`;
+}
+
+function detectChanges(existing: ExistingModel | null, merged: MergedModel): Changes[] {
+  if (!existing) {
+    return [];
+  }
+
+  const changes: Changes[] = [];
+  const epsilon = 0.001;
+
+  const formatValue = (value: unknown): string => {
+    if (typeof value === "number") return formatNumber(value);
+    if (Array.isArray(value)) return `[${value.join(", ")}]`;
+    if (value === undefined) return "(none)";
+    return String(value);
+  };
+
+  const compare = (field: string, oldValue: unknown, newValue: unknown) => {
+    const changed = field.startsWith("cost.")
+      ? (
+          oldValue === undefined && newValue === undefined
+            ? false
+            : oldValue === undefined || newValue === undefined
+              ? true
+              : Math.abs((oldValue as number) - (newValue as number)) > epsilon
+        )
+      : JSON.stringify(oldValue) !== JSON.stringify(newValue);
+
+    if (changed) {
+      changes.push({
+        field,
+        oldValue: formatValue(oldValue),
+        newValue: formatValue(newValue),
+      });
+    }
+  };
+
+  compare("name", existing.name, merged.name);
+  compare("family", existing.family, merged.family);
+  compare("release_date", existing.release_date, merged.release_date);
+  compare("attachment", existing.attachment, merged.attachment);
+  compare("reasoning", existing.reasoning, merged.reasoning);
+  compare("structured_output", existing.structured_output, merged.structured_output);
+  compare("temperature", existing.temperature, merged.temperature);
+  compare("tool_call", existing.tool_call, merged.tool_call);
+  compare("open_weights", existing.open_weights, merged.open_weights);
+  compare("cost.input", existing.cost?.input, merged.cost?.input);
+  compare("cost.output", existing.cost?.output, merged.cost?.output);
+  compare("cost.cache_read", existing.cost?.cache_read, merged.cost?.cache_read);
+  compare("cost.cache_write", existing.cost?.cache_write, merged.cost?.cache_write);
+  compare("limit.context", existing.limit?.context, merged.limit.context);
+  compare("limit.output", existing.limit?.output, merged.limit.output);
+  compare("modalities.input", existing.modalities?.input, merged.modalities.input);
+  compare("modalities.output", existing.modalities?.output, merged.modalities.output);
+
+  return changes;
+}
+
+async function main() {
+  const args = process.argv.slice(2);
+  const dryRun = args.includes("--dry-run");
+  const newOnly = args.includes("--new-only");
+
+  const modelsDir = path.join(import.meta.dirname, "..", "..", "..", "providers", "wandb", "models");
+
+  console.log(`${dryRun ? "[DRY RUN] " : ""}${newOnly ? "[NEW ONLY] " : ""}Fetching WandB models from API...`);
+
+  const res = await fetch(API_ENDPOINT);
+  if (!res.ok) {
+    console.error(`Failed to fetch API: ${res.status} ${res.statusText}`);
+    process.exit(1);
+  }
+
+  const json = await res.json();
+  const parsed = WandbResponse.safeParse(json);
+  if (!parsed.success) {
+    console.error("Invalid API response:", parsed.error.errors);
+    process.exit(1);
+  }
+
+  const apiModels = parsed.data.data;
+  const existingFiles = new Set<string>();
+
+  for await (const file of new Bun.Glob("**/*.toml").scan({ cwd: modelsDir, absolute: false })) {
+    existingFiles.add(file);
+  }
+
+  console.log(`Found ${apiModels.length} models in API, ${existingFiles.size} existing files\n`);
+
+  const apiModelIds = new Set<string>();
+  let created = 0;
+  let updated = 0;
+  let unchanged = 0;
+
+  for (const apiModel of apiModels) {
+    const relativePath = `${apiModel.id}.toml`;
+    const filePath = path.join(modelsDir, relativePath);
+    const dirPath = path.dirname(filePath);
+
+    apiModelIds.add(relativePath);
+
+    const existing = await loadExistingModel(filePath);
+    const merged = mergeModel(apiModel, existing);
+    const tomlContent = formatToml(merged);
+
+    if (existing === null) {
+      created++;
+      if (dryRun) {
+        console.log(`[DRY RUN] Would create: ${relativePath}`);
+        console.log(`  name = "${merged.name}"`);
+        if (merged.family) {
+          console.log(`  family = "${merged.family}"`);
+        }
+        console.log("");
+      } else {
+        await mkdir(dirPath, { recursive: true });
+        await Bun.write(filePath, tomlContent);
+        console.log(`Created: ${relativePath}`);
+      }
+      continue;
+    }
+
+    if (newOnly) {
+      unchanged++;
+      continue;
+    }
+
+    const changes = detectChanges(existing, merged);
+    if (changes.length === 0) {
+      unchanged++;
+      continue;
+    }
+
+    updated++;
+    if (dryRun) {
+      console.log(`[DRY RUN] Would update: ${relativePath}`);
+    } else {
+      await mkdir(dirPath, { recursive: true });
+      await Bun.write(filePath, tomlContent);
+      console.log(`Updated: ${relativePath}`);
+    }
+
+    for (const change of changes) {
+      console.log(`  ${change.field}: ${change.oldValue} → ${change.newValue}`);
+    }
+    console.log("");
+  }
+
+  const orphaned = [...existingFiles].filter((file) => !apiModelIds.has(file));
+  for (const file of orphaned) {
+    console.log(`Warning: Orphaned file (not in API): ${file}`);
+  }
+
+  console.log("");
+  console.log(
+    dryRun
+      ? `Summary: ${created} would be created, ${updated} would be updated, ${unchanged} unchanged, ${orphaned.length} orphaned`
+      : `Summary: ${created} created, ${updated} updated, ${unchanged} unchanged, ${orphaned.length} orphaned`,
+  );
+}
+
+await main();

packages/core/src/family.ts

@@ -96,6 +96,7 @@ export const ModelFamilyValues = [
 
   // NVIDIA Nemotron
   "nemotron",
+  "nemotron-free",
 
   // AWS Titan
   "titan",
@@ -103,6 +104,8 @@ export const ModelFamilyValues = [
 
   // MiniMax
   "minimax",
+  "minimax-m2.5",
+  "minimax-m2.7",
   "minimax-free",
 
   // Hunyuan
@@ -193,6 +196,9 @@ export const ModelFamilyValues = [
 
   // Mimo
   "mimo",
+  "mimo-pro-free",
+  "mimo-omni-free",
+  "mimo-flash-free",
 
   // Clarifai
   "mm-poly",
@@ -284,9 +290,6 @@ export const ModelFamilyValues = [
   // Parakeet
   "parakeet",
 
-  // MiMo
-  "mimo-flash-free",
-
   // NeMo
   "nemoretriever",
 

providers/abacus/models/gemini-3.1-flash-lite-preview.toml

@@ -12,6 +12,8 @@ open_weights = false
 [cost]
 input = 0.25
 output = 1.50
+cache_read = 0.025
+cache_write = 1.00
 
 [limit]
 context = 1_048_576

providers/alibaba-coding-plan/models/MiniMax-M2.5.toml

@@ -18,7 +18,8 @@ cache_read = 0
 cache_write = 0
 
 [limit]
-context = 16_608
+context = 196_608
+input = 196_601
 output = 24_576
 
 [modalities]

providers/amazon-bedrock/models/ai21.jamba-1-5-large-v1:0.toml

@@ -1,22 +0,0 @@
-name = "Jamba 1.5 Large"
-family = "jamba"
-release_date = "2024-08-15"
-last_updated = "2024-08-15"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2024-08"
-tool_call = true
-open_weights = true
-
-[cost]
-input = 2.00
-output = 8.00
-
-[limit]
-context = 256_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/ai21.jamba-1-5-mini-v1:0.toml

@@ -1,22 +0,0 @@
-name = "Jamba 1.5 Mini"
-family = "jamba"
-release_date = "2024-08-15"
-last_updated = "2024-08-15"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2024-08"
-tool_call = true
-open_weights = true
-
-[cost]
-input = 0.20
-output = 0.40
-
-[limit]
-context = 256_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/amazon.titan-text-express-v1.toml

@@ -1,21 +0,0 @@
-name = "Titan Text G1 - Express"
-family = "titan"
-release_date = "2024-12-01"
-last_updated = "2024-12-01"
-attachment = false
-reasoning = false
-temperature = true
-tool_call = true
-open_weights = false
-
-[cost]
-input = 0.20
-output = 0.60
-
-[limit]
-context = 128_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/amazon.titan-text-express-v1:0:8k.toml

@@ -1,21 +0,0 @@
-name = "Titan Text G1 - Express"
-family = "titan"
-release_date = "2024-12-01"
-last_updated = "2024-12-01"
-attachment = false
-reasoning = false
-temperature = true
-tool_call = true
-open_weights = false
-
-[cost]
-input = 0.20
-output = 0.60
-
-[limit]
-context = 128_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/anthropic.claude-3-opus-20240229-v1:0.toml

@@ -1,22 +0,0 @@
-name = "Claude Opus 3"
-family = "claude-opus"
-release_date = "2024-02-29"
-last_updated = "2024-02-29"
-attachment = true
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = true
-open_weights = false
-
-[cost]
-input = 15.00
-output = 75.00
-
-[limit]
-context = 200_000
-output = 4_096
-
-[modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]

providers/amazon-bedrock/models/anthropic.claude-3-sonnet-20240229-v1:0.toml

@@ -1,22 +0,0 @@
-name = "Claude Sonnet 3"
-family = "claude-sonnet"
-release_date = "2024-03-04"
-last_updated = "2024-03-04"
-attachment = true
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = true
-open_weights = false
-
-[cost]
-input = 3.00
-output = 15.00
-
-[limit]
-context = 200_000
-output = 4_096
-
-[modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]

providers/amazon-bedrock/models/anthropic.claude-instant-v1.toml

@@ -1,22 +0,0 @@
-name = "Claude Instant"
-family = "claude"
-release_date = "2023-03-01"
-last_updated = "2023-03-01"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = false
-open_weights = false
-
-[cost]
-input = 0.80
-output = 2.40
-
-[limit]
-context = 100_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/anthropic.claude-opus-4-6-v1.toml

@@ -1,7 +1,7 @@
 name = "Claude Opus 4.6"
 family = "claude-opus"
 release_date = "2026-02-05"
-last_updated = "2026-02-05"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 25.00
 cache_read = 0.50
 cache_write = 6.25
 
-[cost.context_over_200k]
-input = 10.00
-output = 37.50
-cache_read = 1.00
-cache_write = 12.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 128_000
 
 [modalities]

providers/amazon-bedrock/models/anthropic.claude-sonnet-4-6.toml

@@ -1,7 +1,7 @@
 name = "Claude Sonnet 4.6"
 family = "claude-sonnet"
 release_date = "2026-02-17"
-last_updated = "2026-02-17"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 15.00
 cache_read = 0.30
 cache_write = 3.75
 
-[cost.context_over_200k]
-input = 6.00
-output = 22.50
-cache_read = 0.60
-cache_write = 7.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 64_000
 
 [modalities]

providers/amazon-bedrock/models/anthropic.claude-v2.toml

@@ -1,22 +0,0 @@
-name = "Claude 2"
-family = "claude"
-release_date = "2023-07-11"
-last_updated = "2023-07-11"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = false
-open_weights = false
-
-[cost]
-input = 8.00
-output = 24.00
-
-[limit]
-context = 100_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/anthropic.claude-v2:1.toml

@@ -1,22 +0,0 @@
-name = "Claude 2.1"
-family = "claude"
-release_date = "2023-11-21"
-last_updated = "2023-11-21"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = false
-open_weights = false
-
-[cost]
-input = 8.00
-output = 24.00
-
-[limit]
-context = 200_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/cohere.command-light-text-v14.toml

@@ -1,22 +0,0 @@
-name = "Command Light"
-family = "command-light"
-release_date = "2023-11-01"
-last_updated = "2023-11-01"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = false
-open_weights = true
-
-[cost]
-input = 0.30
-output = 0.60
-
-[limit]
-context = 4_096
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/cohere.command-text-v14.toml

@@ -1,22 +0,0 @@
-name = "Command"
-family = "command"
-release_date = "2023-11-01"
-last_updated = "2023-11-01"
-attachment = false
-reasoning = false
-temperature = true
-knowledge = "2023-08"
-tool_call = false
-open_weights = true
-
-[cost]
-input = 1.50
-output = 2.00
-
-[limit]
-context = 4_096
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/deepseek.v3.2.toml

@@ -1,7 +1,7 @@
 name = "DeepSeek-V3.2"
 family = "deepseek"
-release_date = "2026-02-15"
-last_updated = "2026-02-15"
+release_date = "2026-02-06"
+last_updated = "2026-02-06"
 attachment = false
 reasoning = true
 temperature = true

providers/amazon-bedrock/models/eu.anthropic.claude-opus-4-6-v1.toml

@@ -1,7 +1,7 @@
 name = "Claude Opus 4.6 (EU)"
 family = "claude-opus"
 release_date = "2026-02-05"
-last_updated = "2026-02-05"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 25.00
 cache_read = 0.50
 cache_write = 6.25
 
-[cost.context_over_200k]
-input = 10.00
-output = 37.50
-cache_read = 1.00
-cache_write = 12.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 128_000
 
 [modalities]

providers/amazon-bedrock/models/eu.anthropic.claude-sonnet-4-6.toml

@@ -1,7 +1,7 @@
 name = "Claude Sonnet 4.6 (EU)"
 family = "claude-sonnet"
 release_date = "2026-02-17"
-last_updated = "2026-02-17"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 15.00
 cache_read = 0.30
 cache_write = 3.75
 
-[cost.context_over_200k]
-input = 6.00
-output = 22.50
-cache_read = 0.60
-cache_write = 7.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 64_000
 
 [modalities]

providers/amazon-bedrock/models/global.anthropic.claude-opus-4-6-v1.toml

@@ -1,7 +1,7 @@
 name = "Claude Opus 4.6 (Global)"
 family = "claude-opus"
 release_date = "2026-02-05"
-last_updated = "2026-02-05"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 25.00
 cache_read = 0.50
 cache_write = 6.25
 
-[cost.context_over_200k]
-input = 10.00
-output = 37.50
-cache_read = 1.00
-cache_write = 12.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 128_000
 
 [modalities]

providers/amazon-bedrock/models/global.anthropic.claude-sonnet-4-6.toml

@@ -1,7 +1,7 @@
 name = "Claude Sonnet 4.6 (Global)"
 family = "claude-sonnet"
 release_date = "2026-02-17"
-last_updated = "2026-02-17"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 15.00
 cache_read = 0.30
 cache_write = 3.75
 
-[cost.context_over_200k]
-input = 6.00
-output = 22.50
-cache_read = 0.60
-cache_write = 7.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 64_000
 
 [modalities]

providers/amazon-bedrock/models/meta.llama3-1-405b-instruct-v1:0.toml

@@ -1,4 +1,4 @@
-name = "Llama 3 70B Instruct"
+name = "Llama 3.1 405B Instruct"
 family = "llama"
 release_date = "2024-07-23"
 last_updated = "2024-07-23"
@@ -6,16 +6,16 @@ attachment = false
 reasoning = false
 temperature = true
 knowledge = "2023-12"
-tool_call = false
+tool_call = true
 open_weights = true
 
 [cost]
-input = 2.65
-output = 3.50
+input = 2.40
+output = 2.40
 
 [limit]
-context = 8_192
-output = 2_048
+context = 128_000
+output = 4_096
 
 [modalities]
 input = ["text"]

providers/amazon-bedrock/models/mistral.devstral-2-123b.toml

@@ -1,12 +1,12 @@
-name = "Devstral 2 135B"
-family = "mistral"
+name = "Devstral 2 123B"
+family = "devstral"
 release_date = "2026-02-17"
 last_updated = "2026-02-17"
 attachment = false
 reasoning = false
 temperature = true
 tool_call = true
-open_weights = false
+open_weights = true
 
 [cost]
 input = 0.40

providers/amazon-bedrock/models/mistral.magistral-small-2509.toml

@@ -0,0 +1,21 @@
+name = "Magistral Small 1.2"
+family = "magistral"
+release_date = "2025-12-02"
+last_updated = "2025-12-02"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.50
+output = 1.50
+
+[limit]
+context = 128_000
+output = 40_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/amazon-bedrock/models/mistral.ministral-3-3b-instruct.toml

@@ -0,0 +1,21 @@
+name = "Ministral 3 3B"
+family = "ministral"
+release_date = "2025-12-02"
+last_updated = "2025-12-02"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.10
+output = 0.10
+
+[limit]
+context = 256_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/amazon-bedrock/models/mistral.mistral-large-2402-v1:0.toml

@@ -1,21 +0,0 @@
-name = "Mistral Large (24.02)"
-family = "mistral-large"
-release_date = "2024-12-01"
-last_updated = "2024-12-01"
-attachment = false
-reasoning = false
-temperature = true
-tool_call = true
-open_weights = false
-
-[cost]
-input = 0.50
-output = 1.50
-
-[limit]
-context = 128_000
-output = 4_096
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/mistral.mistral-large-3-675b-instruct.toml

@@ -0,0 +1,21 @@
+name = "Mistral Large 3"
+family = "mistral"
+release_date = "2025-12-02"
+last_updated = "2025-12-02"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.50
+output = 1.50
+
+[limit]
+context = 256_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/amazon-bedrock/models/mistral.mixtral-8x7b-instruct-v0:1.toml

@@ -1,22 +0,0 @@
-name = "Mixtral-8x7B-Instruct-v0.1"
-family = "mixtral"
-release_date = "2025-04-01"
-last_updated = "2025-04-01"
-attachment = false
-reasoning = false
-tool_call = false
-structured_output = true
-temperature = true
-open_weights = true
-
-[cost]
-input = 0.70
-output = 0.70
-
-[limit]
-context = 32_000
-output = 32_000
-
-[modalities]
-input = ["text"]
-output = ["text"]

providers/amazon-bedrock/models/mistral.pixtral-large-2502-v1:0.toml

@@ -0,0 +1,21 @@
+name = "Pixtral Large (25.02)"
+family = "mistral"
+release_date = "2025-04-08"
+last_updated = "2025-04-08"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+
+[cost]
+input = 2.00
+output = 6.00
+
+[limit]
+context = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/amazon-bedrock/models/nvidia.nemotron-nano-3-30b.toml

@@ -1,17 +1,16 @@
-name = "Command R"
-family = "command-r"
-release_date = "2024-03-11"
-last_updated = "2024-03-11"
+name = "NVIDIA Nemotron Nano 3 30B"
+family = "nemotron"
+release_date = "2025-12-23"
+last_updated = "2025-12-23"
 attachment = false
-reasoning = false
+reasoning = true
 temperature = true
-knowledge = "2024-04"
 tool_call = true
 open_weights = true
 
 [cost]
-input = 0.50
-output = 1.50
+input = 0.06
+output = 0.24
 
 [limit]
 context = 128_000

providers/amazon-bedrock/models/us.anthropic.claude-opus-4-6-v1.toml

@@ -1,7 +1,7 @@
 name = "Claude Opus 4.6 (US)"
 family = "claude-opus"
 release_date = "2026-02-05"
-last_updated = "2026-02-05"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 25.00
 cache_read = 0.50
 cache_write = 6.25
 
-[cost.context_over_200k]
-input = 10.00
-output = 37.50
-cache_read = 1.00
-cache_write = 12.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 128_000
 
 [modalities]

providers/amazon-bedrock/models/us.anthropic.claude-sonnet-4-6.toml

@@ -1,7 +1,7 @@
 name = "Claude Sonnet 4.6 (US)"
 family = "claude-sonnet"
 release_date = "2026-02-17"
-last_updated = "2026-02-17"
+last_updated = "2026-03-18"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 15.00
 cache_read = 0.30
 cache_write = 3.75
 
-[cost.context_over_200k]
-input = 6.00
-output = 22.50
-cache_read = 0.60
-cache_write = 7.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 64_000
 
 [modalities]

providers/anthropic/models/claude-opus-4-6.toml

@@ -1,7 +1,7 @@
 name = "Claude Opus 4.6"
 family = "claude-opus"
 release_date = "2026-02-05"
-last_updated = "2026-02-05"
+last_updated = "2026-03-13"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 25.00
 cache_read = 0.50
 cache_write = 6.25
 
-[cost.context_over_200k]
-input = 10.00
-output = 37.50
-cache_read = 1.00
-cache_write = 12.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 128_000
 
 [modalities]

providers/anthropic/models/claude-sonnet-4-6.toml

@@ -1,7 +1,7 @@
 name = "Claude Sonnet 4.6"
 family = "claude-sonnet"
 release_date = "2026-02-17"
-last_updated = "2026-02-17"
+last_updated = "2026-03-13"
 attachment = true
 reasoning = true
 temperature = true
@@ -15,14 +15,8 @@ output = 15.00
 cache_read = 0.30
 cache_write = 3.75
 
-[cost.context_over_200k]
-input = 6.00
-output = 22.50
-cache_read = 0.60
-cache_write = 7.50
-
 [limit]
-context = 200_000
+context = 1_000_000
 output = 64_000
 
 [modalities]

providers/azure-cognitive-services/models/gpt-5.4-mini.toml

@@ -0,0 +1,25 @@
+name = "GPT-5.4 Mini"
+family = "gpt-mini"
+release_date = "2026-03-17"
+last_updated = "2026-03-17"
+attachment = true
+reasoning = true
+temperature = false
+knowledge = "2025-08-31"
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.75
+output = 4.50
+cache_read = 0.075
+
+[limit]
+context = 400_000
+input = 272_000
+output = 128_000
+
+[modalities]
+input = ["text", "image", "pdf"]
+output = ["text"]

providers/azure-cognitive-services/models/gpt-5.4-nano.toml

@@ -0,0 +1,25 @@
+name = "GPT-5.4 Nano"
+family = "gpt-nano"
+release_date = "2026-03-17"
+last_updated = "2026-03-17"
+attachment = true
+reasoning = true
+temperature = false
+knowledge = "2025-08-31"
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.20
+output = 1.25
+cache_read = 0.02
+
+[limit]
+context = 400_000
+input = 272_000
+output = 128_000
+
+[modalities]
+input = ["text", "image", "pdf"]
+output = ["text"]

providers/azure/models/cohere-command-r-08-2024.toml

@@ -3,7 +3,7 @@ family = "command-r"
 release_date = "2024-08-30"
 last_updated = "2024-08-30"
 attachment = false
-reasoning = true
+reasoning = false
 temperature = true
 knowledge = "2024-06-01"
 tool_call = true

providers/azure/models/cohere-command-r-plus-08-2024.toml

@@ -3,7 +3,7 @@ family = "command-r"
 release_date = "2024-08-30"
 last_updated = "2024-08-30"
 attachment = false
-reasoning = true
+reasoning = false
 temperature = true
 knowledge = "2024-06-01"
 tool_call = true

providers/azure/models/gpt-5.4-mini.toml

@@ -0,0 +1,25 @@
+name = "GPT-5.4 Mini"
+family = "gpt-mini"
+release_date = "2026-03-17"
+last_updated = "2026-03-17"
+attachment = true
+reasoning = true
+temperature = false
+knowledge = "2025-08-31"
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.75
+output = 4.50
+cache_read = 0.075
+
+[limit]
+context = 400_000
+input = 272_000
+output = 128_000
+
+[modalities]
+input = ["text", "image", "pdf"]
+output = ["text"]

providers/azure/models/gpt-5.4-nano.toml

@@ -0,0 +1,25 @@
+name = "GPT-5.4 Nano"
+family = "gpt-nano"
+release_date = "2026-03-17"
+last_updated = "2026-03-17"
+attachment = true
+reasoning = true
+temperature = false
+knowledge = "2025-08-31"
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.20
+output = 1.25
+cache_read = 0.02
+
+[limit]
+context = 400_000
+input = 272_000
+output = 128_000
+
+[modalities]
+input = ["text", "image", "pdf"]
+output = ["text"]

providers/azure/models/gpt-5.4-pro.toml

@@ -1,4 +1,5 @@
 name = "GPT-5.4 Pro"
+family = "gpt-pro"
 release_date = "2026-03-05"
 last_updated = "2026-03-05"
 attachment = true

providers/azure/models/gpt-5.4.toml

@@ -1,4 +1,5 @@
 name = "GPT-5.4"
+family = "gpt"
 release_date = "2026-03-05"
 last_updated = "2026-03-05"
 attachment = true
@@ -20,5 +21,5 @@ input = 272_000
 output = 128_000
 
 [modalities]
-input = ["text", "image"]
+input = ["text", "image", "pdf"]
 output = ["text"]

providers/azure/models/grok-4-1-fast-non-reasoning.toml

@@ -0,0 +1,24 @@
+name = "Grok 4.1 Fast (Non-Reasoning)"
+family = "grok"
+release_date = "2025-06-27"
+last_updated = "2025-06-27"
+attachment = true
+reasoning = false
+temperature = true
+tool_call = true
+open_weights = false
+status = "beta"
+
+[cost]
+input = 0.20
+output = 0.50
+cache_read = 0.05
+
+[limit]
+context = 128_000
+input = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/azure/models/grok-4-1-fast-reasoning.toml

@@ -0,0 +1,24 @@
+name = "Grok 4.1 Fast (Reasoning)"
+family = "grok"
+release_date = "2025-06-27"
+last_updated = "2025-06-27"
+attachment = true
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = false
+status = "beta"
+
+[cost]
+input = 0.20
+output = 0.50
+cache_read = 0.05
+
+[limit]
+context = 128_000
+input = 128_000
+output = 8_192
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/azure/models/kimi-k2.5.toml

@@ -25,3 +25,5 @@ output = ["text"]
 
 [provider]
 shape = "completions"
+npm = "@ai-sdk/openai-compatible"
+api = "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"

providers/baseten/models/deepseek-ai/DeepSeek-V3-0324.toml

@@ -1,22 +1,22 @@
-name = "DeepSeek-V3-0324"
+name = "DeepSeek V3 0324"
 family = "deepseek"
 release_date = "2025-03-24"
 last_updated = "2025-03-24"
 attachment = false
 reasoning = false
 temperature = true
-knowledge = "2024-10"
+knowledge = "2024-12"
 tool_call = true
 open_weights = true
 
-[cost]
-input = 1.14
-output = 2.75
-
 [limit]
-context = 161_000
-output = 8192
+context = 164_000
+output = 131_000
 
 [modalities]
 input = ["text"]
 output = ["text"]
+
+[cost]
+input = 0.77
+output = 0.77

providers/baseten/models/deepseek-ai/DeepSeek-V3.1.toml

@@ -0,0 +1,21 @@
+name = "DeepSeek V3.1"
+family = "deepseek"
+release_date = "2025-08-25"
+last_updated = "2025-08-25"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[limit]
+context = 164_000
+output = 131_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
+
+[cost]
+input = 0.50
+output = 1.50

providers/baseten/models/deepseek-ai/DeepSeek-V3.2.toml

@@ -1,9 +1,10 @@
 name = "DeepSeek V3.2"
 family = "deepseek"
+status = "deprecated"
 release_date = "2025-12-01"
-last_updated = "2025-12-01"
+last_updated = "2026-03-06"
 attachment = false
-reasoning = false
+reasoning = true
 temperature = true
 knowledge = "2025-10"
 tool_call = true

providers/baseten/models/moonshotai/Kimi-K2-Instruct-0905.toml

@@ -1,7 +1,8 @@
 name = "Kimi K2 Instruct 0905"
 family = "kimi"
+status = "deprecated"
 release_date = "2025-09-05"
-last_updated = "2025-09-05"
+last_updated = "2026-03-06"
 attachment = false
 reasoning = false
 temperature = true

providers/baseten/models/moonshotai/Kimi-K2-Thinking.toml

@@ -1,7 +1,8 @@
 name = "Kimi K2 Thinking"
 family = "kimi-thinking"
+status = "deprecated"
 release_date = "2025-11-06"
-last_updated = "2025-11-06"
+last_updated = "2026-03-06"
 attachment = false
 reasoning = true
 temperature = true

providers/baseten/models/nvidia/Nemotron-3-Super.toml

@@ -0,0 +1,25 @@
+name = "Nemotron 3 Super"
+family = "nemotron"
+release_date = "2026-03-11"
+last_updated = "2026-03-11"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+knowledge = "2026-02"
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.30
+output = 0.75
+
+[limit]
+context = 262144
+output = 32678
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/baseten/models/openai/gpt-oss-120b.toml

@@ -1,20 +1,22 @@
-name = "OpenAI: gpt-oss-120b (exacto)"
+name = "GPT OSS 120B"
+family = "gpt-oss"
 release_date = "2025-08-05"
 last_updated = "2025-08-05"
 attachment = false
 reasoning = true
 temperature = true
+knowledge = "2025-08"
 tool_call = true
 open_weights = true
 
-[cost]
-input = 0.039
-output = 0.19
-
 [limit]
-context = 131072
-output = 26215
+context = 128_000
+output = 128_000
 
 [modalities]
 input = ["text"]
 output = ["text"]
+
+[cost]
+input = 0.10
+output = 0.50

providers/chutes/models/MiniMaxAI/MiniMax-M2.5-TEE.toml

@@ -10,8 +10,9 @@ structured_output = true
 open_weights = true
 
 [cost]
-input = 0.15
-output = 0.6
+input = 0.30
+output = 1.10
+cache_read = 0.15
 
 [limit]
 context = 196_608

providers/chutes/models/Qwen/Qwen3.5-397B-A17B-TEE.toml

@@ -10,9 +10,9 @@ structured_output = true
 open_weights = true
 
 [cost]
-input = 0.30
-output = 1.20
-cache_read = 0.15
+input = 0.39
+output = 2.34
+cache_read = 0.195
 
 [limit]
 context = 262_144

providers/chutes/models/deepseek-ai/DeepSeek-V3.2-TEE.toml

@@ -10,12 +10,12 @@ structured_output = true
 open_weights = true
 
 [cost]
-input = 0.25
-output = 0.38
-cache_read = 0.125
+input = 0.28
+output = 0.42
+cache_read = 0.14
 
 [limit]
-context = 163_840
+context = 131_072
 output = 65_536
 
 [modalities]

providers/chutes/models/zai-org/GLM-4.6-TEE.toml

@@ -10,8 +10,9 @@ structured_output = true
 open_weights = true
 
 [cost]
-input = 0.35
-output = 1.50
+input = 0.40
+output = 1.70
+cache_read = 0.20
 
 [limit]
 context = 202_752

providers/chutes/models/zai-org/GLM-4.6V.toml

@@ -12,6 +12,7 @@ open_weights = true
 [cost]
 input = 0.30
 output = 0.90
+cache_read = 0.15
 
 [limit]
 context = 131_072

providers/chutes/models/zai-org/GLM-5-TEE.toml

@@ -10,8 +10,9 @@ structured_output = true
 open_weights = true
 
 [cost]
-input = 0.750
-output = 2.50
+input = 0.95
+output = 3.15
+cache_read = 0.475
 
 [limit]
 context = 202_752

providers/chutes/models/zai-org/GLM-5-Turbo.toml

@@ -0,0 +1,26 @@
+name = "GLM 5 Turbo"
+family = "glm"
+release_date = "2026-03-11"
+last_updated = "2026-03-11"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = true
+
+[cost]
+input = 0.49
+output = 1.96
+cache_read = 0.245
+
+[limit]
+context = 202_752
+output = 65_535
+
+[modalities]
+input = ["text"]
+output = ["text"]
+
+[interleaved]
+field = "reasoning_content"

providers/cloudflare-ai-gateway/models/workers-ai/@cf/moonshotai/kimi-k2.5.toml

@@ -0,0 +1,27 @@
+name = "Kimi K2.5"
+family = "kimi"
+release_date = "2026-01-27"
+last_updated = "2026-01-27"
+attachment = true
+reasoning = true
+structured_output = true
+temperature = true
+tool_call = true
+knowledge = "2025-01"
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.60
+output = 3.00
+cache_read = 0.10
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/cloudflare-ai-gateway/models/workers-ai/@cf/nvidia/nemotron-3-120b-a12b.toml

@@ -0,0 +1,24 @@
+name = "Nemotron 3 Super 120B"
+family = "nemotron"
+release_date = "2026-03-11"
+last_updated = "2026-03-11"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.50
+output = 1.50
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/cloudflare-ai-gateway/models/workers-ai/@cf/zai-org/glm-4.7-flash.toml

@@ -0,0 +1,22 @@
+name = "GLM-4.7-Flash"
+family = "glm-flash"
+release_date = "2026-01-19"
+last_updated = "2026-01-19"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+knowledge = "2025-04"
+open_weights = true
+
+[cost]
+input = 0.06
+output = 0.40
+
+[limit]
+context = 131_072
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/cloudflare-workers-ai/models/@cf/moonshotai/kimi-k2.5.toml

@@ -0,0 +1,27 @@
+name = "Kimi K2.5"
+family = "kimi"
+release_date = "2026-01-27"
+last_updated = "2026-01-27"
+attachment = true
+reasoning = true
+structured_output = true
+temperature = true
+tool_call = true
+knowledge = "2025-01"
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.60
+output = 3.00
+cache_read = 0.10
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/cloudflare-workers-ai/models/@cf/nvidia/nemotron-3-120b-a12b.toml

@@ -0,0 +1,24 @@
+name = "Nemotron 3 Super 120B"
+family = "nemotron"
+release_date = "2026-03-11"
+last_updated = "2026-03-11"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.50
+output = 1.50
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/cohere/models/c4ai-aya-expanse-32b.toml

@@ -4,7 +4,7 @@ last_updated = "2024-10-24"
 attachment = false
 reasoning = false
 temperature = true
-tool_call = true
+tool_call = false
 open_weights = true
 
 [limit]

providers/cohere/models/c4ai-aya-expanse-8b.toml

@@ -4,7 +4,7 @@ last_updated = "2024-10-24"
 attachment = false
 reasoning = false
 temperature = true
-tool_call = true
+tool_call = false
 open_weights = true
 
 [limit]

providers/cohere/models/c4ai-aya-vision-32b.toml

@@ -4,7 +4,7 @@ last_updated = "2025-05-14"
 attachment = true
 reasoning = false
 temperature = true
-tool_call = true
+tool_call = false
 open_weights = true
 
 [limit]

providers/cohere/models/c4ai-aya-vision-8b.toml

@@ -4,7 +4,7 @@ last_updated = "2025-05-14"
 attachment = true
 reasoning = false
 temperature = true
-tool_call = true
+tool_call = false
 open_weights = true
 
 [limit]

providers/cohere/models/command-a-03-2025.toml

@@ -3,7 +3,7 @@ family = "command-a"
 release_date = "2025-03-13"
 last_updated = "2025-03-13"
 attachment = false
-reasoning = true
+reasoning = false
 temperature = true
 knowledge = "2024-06-01"
 tool_call = true

providers/cohere/models/command-r-08-2024.toml

@@ -3,7 +3,7 @@ family = "command-r"
 release_date = "2024-08-30"
 last_updated = "2024-08-30"
 attachment = false
-reasoning = true
+reasoning = false
 temperature = true
 knowledge = "2024-06-01"
 tool_call = true

providers/cohere/models/command-r-plus-08-2024.toml

@@ -3,7 +3,7 @@ family = "command-r"
 release_date = "2024-08-30"
 last_updated = "2024-08-30"
 attachment = false
-reasoning = true
+reasoning = false
 temperature = true
 knowledge = "2024-06-01"
 tool_call = true

providers/deepinfra/models/meta-llama/Llama-3.1-70B-Instruct-Turbo.toml

@@ -1,21 +1,19 @@
-name = "Llama 3 8B Instruct"
+name = "Llama 3.1 70B Turbo"
 family = "llama"
 release_date = "2024-07-23"
 last_updated = "2024-07-23"
 attachment = false
 reasoning = false
-temperature = true
-knowledge = "2023-03"
-tool_call = false
+tool_call = true
 open_weights = true
 
 [cost]
-input = 0.30
-output = 0.60
+input = 0.40
+output = 0.40
 
 [limit]
-context = 8_192
-output = 2_048
+context = 131_072
+output = 16_384
 
 [modalities]
 input = ["text"]

providers/deepinfra/models/meta-llama/Llama-3.1-70B-Instruct.toml

@@ -0,0 +1,20 @@
+name = "Llama 3.1 70B"
+family = "llama"
+release_date = "2024-07-23"
+last_updated = "2024-07-23"
+attachment = false
+reasoning = false
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.40
+output = 0.40
+
+[limit]
+context = 131_072
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/deepinfra/models/meta-llama/Llama-3.1-8B-Instruct-Turbo.toml

@@ -0,0 +1,20 @@
+name = "Llama 3.1 8B Turbo"
+family = "llama"
+release_date = "2024-07-23"
+last_updated = "2024-07-23"
+attachment = false
+reasoning = false
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.02
+output = 0.03
+
+[limit]
+context = 131_072
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/deepinfra/models/meta-llama/Llama-3.1-8B-Instruct.toml

@@ -0,0 +1,20 @@
+name = "Llama 3.1 8B"
+family = "llama"
+release_date = "2024-07-23"
+last_updated = "2024-07-23"
+attachment = false
+reasoning = false
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.02
+output = 0.05
+
+[limit]
+context = 131_072
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/deepinfra/models/meta-llama/Llama-3.3-70B-Instruct-Turbo.toml

@@ -1,19 +1,19 @@
-name = "Meta: Llama 3.3 70B Instruct (free)"
+name = "Llama 3.3 70B Turbo"
+family = "llama"
 release_date = "2024-12-06"
 last_updated = "2024-12-06"
 attachment = false
 reasoning = false
-temperature = true
 tool_call = true
 open_weights = true
 
 [cost]
-input = 0
-output = 0
+input = 0.10
+output = 0.32
 
 [limit]
-context = 128000
-output = 128000
+context = 131_072
+output = 16_384
 
 [modalities]
 input = ["text"]

providers/deepinfra/models/meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8.toml

@@ -0,0 +1,20 @@
+name = "Llama 4 Maverick 17B FP8"
+family = "llama"
+release_date = "2025-04-05"
+last_updated = "2025-04-05"
+attachment = false
+reasoning = false
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.15
+output = 0.60
+
+[limit]
+context = 1_000_000
+output = 16_384
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/deepinfra/models/meta-llama/Llama-4-Scout-17B-16E-Instruct.toml

@@ -0,0 +1,20 @@
+name = "Llama 4 Scout 17B"
+family = "llama"
+release_date = "2025-04-05"
+last_updated = "2025-04-05"
+attachment = false
+reasoning = false
+tool_call = true
+open_weights = true
+
+[cost]
+input = 0.08
+output = 0.30
+
+[limit]
+context = 10_000_000
+output = 16_384
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]

providers/dinference/logo.svg

@@ -0,0 +1 @@
+<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="2" class="lucide lucide-terminal h-5 w-5 text-primary"><path d="m4 17 6-6-6-6M12 19h8"/></svg>

providers/dinference/models/glm-4.7.toml

@@ -0,0 +1,25 @@
+name = "GLM-4.7"
+family = "glm"
+release_date = "2025-12"
+last_updated = "2025-12"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+knowledge = "2025-04"
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.45
+output = 1.65
+
+[limit]
+context = 200_000
+output = 128_000
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/dinference/models/glm-5.toml

@@ -0,0 +1,25 @@
+name = "GLM-5"
+family = "glm"
+release_date = "2026-02"
+last_updated = "2026-02"
+attachment = false
+reasoning = true
+temperature = true
+tool_call = true
+knowledge = "2025-04"
+open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[cost]
+input = 0.75
+output = 2.40
+
+[limit]
+context = 200_000
+output = 128_000
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/dinference/models/gpt-oss-120b.toml

@@ -1,6 +1,6 @@
-name = "MoonshotAI: Kimi K2 0905 (exacto)"
-release_date = "2025-09-05"
-last_updated = "2025-09-05"
+name = "GPT OSS 120B"
+release_date = "2025-08"
+last_updated = "2025-08"
 attachment = false
 reasoning = false
 temperature = true
@@ -8,12 +8,12 @@ tool_call = true
 open_weights = true
 
 [cost]
-input = 0.6
-output = 2.5
+input = 0.0675
+output = 0.27
 
 [limit]
-context = 262144
-output = 52429
+context = 131_072
+output = 32_768
 
 [modalities]
 input = ["text"]

providers/dinference/provider.toml

@@ -0,0 +1,5 @@
+name = "DInference"
+env = ["DINFERENCE_API_KEY"]
+npm = "@ai-sdk/openai-compatible"
+doc = "https://dinference.com"
+api = "https://api.dinference.com/v1"

providers/firmware/provider.toml

@@ -2,5 +2,5 @@ name = "Firmware"
 # Token for Firmware's OpenAI-compatible proxy
 env = ["FIRMWARE_API_KEY"]
 npm = "@ai-sdk/openai-compatible"
-api = "https://app.firmware.ai/api/v1"
-doc = "https://docs.firmware.ai"
+api = "https://app.frogbot.ai/api/v1"
+doc = "https://docs.frogbot.ai"

providers/github-copilot/models/claude-haiku-4.5.toml

@@ -14,8 +14,9 @@ input = 0
 output = 0
 
 [limit]
-context = 128_000
+context = 144_000
 output = 32_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/claude-opus-4.5.toml

@@ -14,8 +14,9 @@ input = 0
 output = 0
 
 [limit]
-context = 128_000
+context = 160_000
 output = 32_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/claude-opus-4.6.toml

@@ -14,8 +14,9 @@ input = 0
 output = 0
 
 [limit]
-context = 128_000
+context = 144_000
 output = 64_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/claude-sonnet-4.5.toml

@@ -14,8 +14,9 @@ input = 0
 output = 0
 
 [limit]
-context = 128_000
+context = 144_000
 output = 32_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/claude-sonnet-4.6.toml

@@ -13,8 +13,9 @@ input = 0
 output = 0
 
 [limit]
-context = 128_000
+context = 200_000
 output = 32_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/claude-sonnet-4.toml

@@ -14,8 +14,9 @@ input = 0
 output = 0
 
 [limit]
-context = 128_000
+context = 216_000
 output = 16_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/gemini-2.5-pro.toml

@@ -16,6 +16,7 @@ output = 0
 [limit]
 context = 128_000
 output = 64_000
+input = 128_000
 
 [modalities]
 input = ["text", "image", "audio", "video"]

providers/github-copilot/models/gemini-3-flash-preview.toml

@@ -17,6 +17,7 @@ output = 0
 [limit]
 context = 128_000
 output = 64_000
+input = 128_000
 
 [modalities]
 input = ["text", "image", "audio", "video"]

providers/github-copilot/models/gemini-3-pro-preview.toml

@@ -17,6 +17,7 @@ output = 0
 [limit]
 context = 128_000
 output = 64_000
+input = 128_000
 
 [modalities]
 input = ["text", "image", "audio", "video"]

providers/github-copilot/models/gemini-3.1-pro-preview.toml

@@ -17,6 +17,7 @@ output = 0
 [limit]
 context = 128_000
 output = 64_000
+input = 128_000
 
 [modalities]
 input = ["text", "image"]

providers/github-copilot/models/gpt-4.1.toml

@@ -14,8 +14,9 @@ input = 0
 output = 0
 
 [limit]
-context = 64_000
+context = 128_000
 output = 16_384
+input = 64_000
 
 [modalities]
 input = ["text", "image"]