@codilore/llm 1.15.13

package/test/provider/gemini.test.ts

@@ -0,0 +1,476 @@
+import { describe, expect } from "bun:test"
+import { Effect } from "effect"
+import { LLM, LLMError, Message, ToolCallPart, Usage } from "../../src"
+import { Auth, LLMClient } from "../../src/route"
+import * as Gemini from "../../src/protocols/gemini"
+import { it } from "../lib/effect"
+import { fixedResponse } from "../lib/http"
+import { sseEvents, sseRaw } from "../lib/sse"
+
+const model = Gemini.route
+  .with({
+    endpoint: { baseURL: "https://generativelanguage.test/v1beta/" },
+    auth: Auth.header("x-goog-api-key", "test"),
+  })
+  .model({ id: "gemini-2.5-flash" })
+
+const request = LLM.request({
+  id: "req_1",
+  model,
+  system: "You are concise.",
+  prompt: "Say hello.",
+  generation: { maxTokens: 20, temperature: 0 },
+})
+
+describe("Gemini route", () => {
+  it.effect("prepares Gemini target", () =>
+    Effect.gen(function* () {
+      const prepared = yield* LLMClient.prepare(request)
+
+      expect(prepared.body).toEqual({
+        contents: [{ role: "user", parts: [{ text: "Say hello." }] }],
+        systemInstruction: { parts: [{ text: "You are concise." }] },
+        generationConfig: { maxOutputTokens: 20, temperature: 0 },
+      })
+    }),
+  )
+
+  it.effect("lowers chronological system updates to wrapped user text in order", () =>
+    Effect.gen(function* () {
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
+        LLM.request({
+          model,
+          messages: [Message.user("Before."), Message.system("Update."), Message.assistant("After.")],
+        }),
+      )
+
+      expect(prepared.body.contents).toEqual([
+        { role: "user", parts: [{ text: "Before." }, { text: "<system-update>\nUpdate.\n</system-update>" }] },
+        { role: "model", parts: [{ text: "After." }] },
+      ])
+    }),
+  )
+
+  it.effect("prepares multimodal user input and tool history", () =>
+    Effect.gen(function* () {
+      const prepared = yield* LLMClient.prepare(
+        LLM.request({
+          id: "req_tool_result",
+          model,
+          tools: [
+            {
+              name: "lookup",
+              description: "Lookup data",
+              inputSchema: { type: "object", properties: { query: { type: "string" } } },
+            },
+          ],
+          toolChoice: { type: "tool", name: "lookup" },
+          messages: [
+            Message.user([
+              { type: "text", text: "What is in this image?" },
+              { type: "media", mediaType: "image/png", data: "AAECAw==" },
+            ]),
+            Message.assistant([ToolCallPart.make({ id: "call_1", name: "lookup", input: { query: "weather" } })]),
+            Message.tool({ id: "call_1", name: "lookup", result: { forecast: "sunny" } }),
+          ],
+        }),
+      )
+
+      expect(prepared.body).toEqual({
+        contents: [
+          {
+            role: "user",
+            parts: [{ text: "What is in this image?" }, { inlineData: { mimeType: "image/png", data: "AAECAw==" } }],
+          },
+          {
+            role: "model",
+            parts: [{ functionCall: { name: "lookup", args: { query: "weather" } } }],
+          },
+          {
+            role: "user",
+            parts: [
+              { functionResponse: { name: "lookup", response: { name: "lookup", content: '{"forecast":"sunny"}' } } },
+            ],
+          },
+        ],
+        tools: [
+          {
+            functionDeclarations: [
+              {
+                name: "lookup",
+                description: "Lookup data",
+                parameters: { type: "object", properties: { query: { type: "string" } } },
+              },
+            ],
+          },
+        ],
+        toolConfig: { functionCallingConfig: { mode: "ANY", allowedFunctionNames: ["lookup"] } },
+      })
+    }),
+  )
+
+  it.effect("omits tools when tool choice is none", () =>
+    Effect.gen(function* () {
+      const prepared = yield* LLMClient.prepare(
+        LLM.request({
+          id: "req_no_tools",
+          model,
+          prompt: "Say hello.",
+          tools: [{ name: "lookup", description: "Lookup data", inputSchema: { type: "object" } }],
+          toolChoice: { type: "none" },
+        }),
+      )
+
+      expect(prepared.body).toEqual({
+        contents: [{ role: "user", parts: [{ text: "Say hello." }] }],
+      })
+    }),
+  )
+
+  it.effect("sanitizes integer enums, dangling required, untyped arrays, and scalar object keys", () =>
+    Effect.gen(function* () {
+      const prepared = yield* LLMClient.prepare(
+        LLM.request({
+          id: "req_schema_patch",
+          model,
+          prompt: "Use the tool.",
+          tools: [
+            {
+              name: "lookup",
+              description: "Lookup data",
+              inputSchema: {
+                type: "object",
+                required: ["status", "missing"],
+                properties: {
+                  status: { type: "integer", enum: [1, 2] },
+                  tags: { type: "array" },
+                  name: { type: "string", properties: { ignored: { type: "string" } }, required: ["ignored"] },
+                },
+              },
+            },
+          ],
+        }),
+      )
+
+      expect(prepared.body).toMatchObject({
+        tools: [
+          {
+            functionDeclarations: [
+              {
+                parameters: {
+                  type: "object",
+                  required: ["status"],
+                  properties: {
+                    status: { type: "string", enum: ["1", "2"] },
+                    tags: { type: "array", items: { type: "string" } },
+                    name: { type: "string" },
+                  },
+                },
+              },
+            ],
+          },
+        ],
+      })
+    }),
+  )
+
+  it.effect("parses text, reasoning, and usage stream fixtures", () =>
+    Effect.gen(function* () {
+      const body = sseEvents(
+        {
+          candidates: [
+            {
+              content: { role: "model", parts: [{ text: "thinking", thought: true }] },
+            },
+          ],
+        },
+        {
+          candidates: [
+            {
+              content: { role: "model", parts: [{ text: "Hello" }] },
+            },
+          ],
+        },
+        {
+          candidates: [
+            {
+              content: { role: "model", parts: [{ text: "!" }] },
+              finishReason: "STOP",
+            },
+          ],
+        },
+        {
+          usageMetadata: {
+            promptTokenCount: 5,
+            candidatesTokenCount: 2,
+            totalTokenCount: 7,
+            thoughtsTokenCount: 1,
+            cachedContentTokenCount: 1,
+          },
+        },
+      )
+      const response = yield* LLMClient.generate(request).pipe(Effect.provide(fixedResponse(body)))
+
+      expect(response.text).toBe("Hello!")
+      expect(response.reasoning).toBe("thinking")
+      expect(response.usage).toMatchObject({
+        inputTokens: 5,
+        outputTokens: 3,
+        nonCachedInputTokens: 4,
+        cacheReadInputTokens: 1,
+        reasoningTokens: 1,
+        totalTokens: 7,
+      })
+      const usage = new Usage({
+        inputTokens: 5,
+        outputTokens: 3,
+        nonCachedInputTokens: 4,
+        cacheReadInputTokens: 1,
+        reasoningTokens: 1,
+        totalTokens: 7,
+        providerMetadata: {
+          google: {
+            promptTokenCount: 5,
+            candidatesTokenCount: 2,
+            totalTokenCount: 7,
+            thoughtsTokenCount: 1,
+            cachedContentTokenCount: 1,
+          },
+        },
+      })
+      expect(response.events).toEqual([
+        { type: "step-start", index: 0 },
+        { type: "reasoning-start", id: "reasoning-0" },
+        { type: "reasoning-delta", id: "reasoning-0", text: "thinking" },
+        { type: "text-start", id: "text-0" },
+        { type: "text-delta", id: "text-0", text: "Hello" },
+        { type: "text-delta", id: "text-0", text: "!" },
+        { type: "reasoning-end", id: "reasoning-0" },
+        { type: "text-end", id: "text-0" },
+        { type: "step-finish", index: 0, reason: "stop", usage, providerMetadata: undefined },
+        {
+          type: "finish",
+          reason: "stop",
+          usage,
+        },
+      ])
+    }),
+  )
+
+  it.effect("preserves thoughtSignature for reasoning and tool-call continuation", () =>
+    Effect.gen(function* () {
+      const body = sseEvents({
+        candidates: [
+          {
+            content: {
+              role: "model",
+              parts: [
+                { text: "thinking", thought: true },
+                { text: "", thought: true, thoughtSignature: "thought_sig" },
+                { functionCall: { name: "lookup", args: { query: "weather" } }, thoughtSignature: "tool_sig" },
+              ],
+            },
+            finishReason: "STOP",
+          },
+        ],
+      })
+      const response = yield* LLMClient.generate(
+        LLM.updateRequest(request, {
+          tools: [{ name: "lookup", description: "Lookup data", inputSchema: { type: "object" } }],
+        }),
+      ).pipe(Effect.provide(fixedResponse(body)))
+      const reasoning = response.events.find((event) => event.type === "reasoning-start")
+      const reasoningEnd = response.events.find((event) => event.type === "reasoning-end")
+      const toolCall = response.events.find((event) => event.type === "tool-call")
+
+      expect(reasoning).toEqual({
+        type: "reasoning-start",
+        id: "reasoning-0",
+        providerMetadata: undefined,
+      })
+      expect(reasoningEnd).toEqual({
+        type: "reasoning-end",
+        id: "reasoning-0",
+        providerMetadata: { google: { thoughtSignature: "thought_sig" } },
+      })
+      expect(toolCall).toMatchObject({ providerMetadata: { google: { thoughtSignature: "tool_sig" } } })
+
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
+        LLM.request({
+          model,
+          messages: [
+            Message.assistant([
+              { type: "reasoning", text: "thinking", providerMetadata: reasoningEnd?.providerMetadata },
+              ToolCallPart.make({
+                id: "tool_0",
+                name: "lookup",
+                input: { query: "weather" },
+                providerMetadata: toolCall?.providerMetadata,
+              }),
+            ]),
+          ],
+        }),
+      )
+      expect(prepared.body.contents).toEqual([
+        {
+          role: "model",
+          parts: [
+            { text: "thinking", thought: true, thoughtSignature: "thought_sig" },
+            { functionCall: { name: "lookup", args: { query: "weather" } }, thoughtSignature: "tool_sig" },
+          ],
+        },
+      ])
+    }),
+  )
+
+  it.effect("emits streamed tool calls and maps finish reason", () =>
+    Effect.gen(function* () {
+      const body = sseEvents({
+        candidates: [
+          {
+            content: {
+              role: "model",
+              parts: [{ functionCall: { name: "lookup", args: { query: "weather" } } }],
+            },
+            finishReason: "STOP",
+          },
+        ],
+        usageMetadata: { promptTokenCount: 5, candidatesTokenCount: 1 },
+      })
+      const response = yield* LLMClient.generate(
+        LLM.updateRequest(request, {
+          tools: [{ name: "lookup", description: "Lookup data", inputSchema: { type: "object" } }],
+        }),
+      ).pipe(Effect.provide(fixedResponse(body)))
+      const usage = new Usage({
+        inputTokens: 5,
+        outputTokens: 1,
+        nonCachedInputTokens: 5,
+        cacheReadInputTokens: undefined,
+        reasoningTokens: undefined,
+        totalTokens: 6,
+        providerMetadata: { google: { promptTokenCount: 5, candidatesTokenCount: 1 } },
+      })
+
+      expect(response.toolCalls).toEqual([
+        {
+          type: "tool-call",
+          id: "tool_0",
+          name: "lookup",
+          input: { query: "weather" },
+          providerExecuted: undefined,
+          providerMetadata: undefined,
+        },
+      ])
+      expect(response.events).toEqual([
+        { type: "step-start", index: 0 },
+        {
+          type: "tool-call",
+          id: "tool_0",
+          name: "lookup",
+          input: { query: "weather" },
+          providerExecuted: undefined,
+          providerMetadata: undefined,
+        },
+        { type: "step-finish", index: 0, reason: "tool-calls", usage, providerMetadata: undefined },
+        {
+          type: "finish",
+          reason: "tool-calls",
+          usage,
+        },
+      ])
+    }),
+  )
+
+  it.effect("assigns unique ids to multiple streamed tool calls", () =>
+    Effect.gen(function* () {
+      const body = sseEvents({
+        candidates: [
+          {
+            content: {
+              role: "model",
+              parts: [
+                { functionCall: { name: "lookup", args: { query: "weather" } } },
+                { functionCall: { name: "lookup", args: { query: "news" } } },
+              ],
+            },
+            finishReason: "STOP",
+          },
+        ],
+      })
+      const response = yield* LLMClient.generate(
+        LLM.updateRequest(request, {
+          tools: [{ name: "lookup", description: "Lookup data", inputSchema: { type: "object" } }],
+        }),
+      ).pipe(Effect.provide(fixedResponse(body)))
+
+      expect(response.toolCalls).toEqual([
+        { type: "tool-call", id: "tool_0", name: "lookup", input: { query: "weather" } },
+        { type: "tool-call", id: "tool_1", name: "lookup", input: { query: "news" } },
+      ])
+      expect(response.events.at(-1)).toMatchObject({ type: "finish", reason: "tool-calls" })
+    }),
+  )
+
+  it.effect("maps length and content-filter finish reasons", () =>
+    Effect.gen(function* () {
+      const length = yield* LLMClient.generate(request).pipe(
+        Effect.provide(
+          fixedResponse(
+            sseEvents({ candidates: [{ content: { role: "model", parts: [] }, finishReason: "MAX_TOKENS" }] }),
+          ),
+        ),
+      )
+      const filtered = yield* LLMClient.generate(request).pipe(
+        Effect.provide(
+          fixedResponse(sseEvents({ candidates: [{ content: { role: "model", parts: [] }, finishReason: "SAFETY" }] })),
+        ),
+      )
+
+      expect(length.events.map((event) => event.type)).toEqual(["step-start", "step-finish", "finish"])
+      expect(length.events.at(-1)).toMatchObject({ type: "finish", reason: "length" })
+      expect(filtered.events.map((event) => event.type)).toEqual(["step-start", "step-finish", "finish"])
+      expect(filtered.events.at(-1)).toMatchObject({ type: "finish", reason: "content-filter" })
+    }),
+  )
+
+  it.effect("leaves total usage undefined when component counts are missing", () =>
+    Effect.gen(function* () {
+      const response = yield* LLMClient.generate(request).pipe(
+        Effect.provide(fixedResponse(sseEvents({ usageMetadata: { thoughtsTokenCount: 1 } }))),
+      )
+
+      expect(response.usage).toMatchObject({ reasoningTokens: 1 })
+      expect(response.usage?.totalTokens).toBeUndefined()
+    }),
+  )
+
+  it.effect("fails invalid stream events", () =>
+    Effect.gen(function* () {
+      const error = yield* LLMClient.generate(request).pipe(
+        Effect.provide(fixedResponse(sseRaw("data: {not json}"))),
+        Effect.flip,
+      )
+
+      expect(error).toBeInstanceOf(LLMError)
+      expect(error.reason).toMatchObject({ _tag: "InvalidProviderOutput" })
+      expect(error.message).toContain("Invalid google/gemini stream event")
+    }),
+  )
+
+  it.effect("rejects unsupported assistant media content", () =>
+    Effect.gen(function* () {
+      const error = yield* LLMClient.prepare(
+        LLM.request({
+          id: "req_media",
+          model,
+          messages: [Message.assistant({ type: "media", mediaType: "image/png", data: "AAECAw==" })],
+        }),
+      ).pipe(Effect.flip)
+
+      expect(error.message).toContain(
+        "Gemini assistant messages only support text, reasoning, and tool-call content for now",
+      )
+    }),
+  )
+})

package/test/provider/golden.recorded.test.ts

@@ -0,0 +1,219 @@
+import { Redactor } from "@codilore/http-recorder"
+import * as Anthropic from "../../src/providers/anthropic"
+import { CloudflareAIGateway, CloudflareWorkersAI } from "../../src/providers/cloudflare"
+import * as Google from "../../src/providers/google"
+import * as OpenAI from "../../src/providers/openai"
+import * as OpenAICompatible from "../../src/providers/openai-compatible"
+import * as OpenRouter from "../../src/providers/openrouter"
+import * as XAI from "../../src/providers/xai"
+import { describeRecordedGoldenScenarios } from "../recorded-golden"
+
+const openAI = OpenAI.configure({
+  apiKey: process.env.OPENAI_API_KEY ?? "fixture",
+})
+const openAIChat = openAI.chat("gpt-4o-mini")
+const openAIResponses = openAI.responses("gpt-5.5")
+const openAIResponsesWebSocket = openAI.responsesWebSocket("gpt-4.1-mini")
+const anthropic = Anthropic.configure({
+  apiKey: process.env.ANTHROPIC_API_KEY ?? "fixture",
+})
+const anthropicHaiku = anthropic.model("claude-haiku-4-5-20251001")
+const anthropicOpus = anthropic.model("claude-opus-4-7")
+const google = Google.configure({ apiKey: process.env.GOOGLE_GENERATIVE_AI_API_KEY ?? "fixture" })
+const gemini = google.model("gemini-2.5-flash")
+const xai = XAI.configure({ apiKey: process.env.XAI_API_KEY ?? "fixture" })
+const xaiBasic = xai.model("grok-3-mini")
+const xaiFlagship = xai.model("grok-4.3")
+const cloudflareAIGateway = CloudflareAIGateway.configure({
+  accountId: process.env.CLOUDFLARE_ACCOUNT_ID ?? "fixture-account",
+  gatewayId:
+    process.env.CLOUDFLARE_GATEWAY_ID && process.env.CLOUDFLARE_GATEWAY_ID !== process.env.CLOUDFLARE_ACCOUNT_ID
+      ? process.env.CLOUDFLARE_GATEWAY_ID
+      : undefined,
+  gatewayApiKey: process.env.CLOUDFLARE_API_TOKEN ?? "fixture",
+})
+const cloudflareWorkers = CloudflareWorkersAI.configure({
+  accountId: process.env.CLOUDFLARE_ACCOUNT_ID ?? "fixture-account",
+  apiKey: process.env.CLOUDFLARE_API_KEY ?? "fixture",
+})
+const cloudflareAIGatewayWorkers = cloudflareAIGateway.model("workers-ai/@cf/meta/llama-3.1-8b-instruct")
+const cloudflareAIGatewayWorkersTools = cloudflareAIGateway.model("workers-ai/@cf/openai/gpt-oss-20b")
+const cloudflareWorkersAI = cloudflareWorkers.model("@cf/meta/llama-3.1-8b-instruct")
+const cloudflareWorkersAITools = cloudflareWorkers.model("@cf/openai/gpt-oss-20b")
+const deepseek = OpenAICompatible.deepseek
+  .configure({ apiKey: process.env.DEEPSEEK_API_KEY ?? "fixture" })
+  .model("deepseek-chat")
+const together = OpenAICompatible.togetherai
+  .configure({
+    apiKey: process.env.TOGETHER_AI_API_KEY ?? "fixture",
+  })
+  .model("meta-llama/Llama-3.3-70B-Instruct-Turbo")
+const groq = OpenAICompatible.groq
+  .configure({ apiKey: process.env.GROQ_API_KEY ?? "fixture" })
+  .model("llama-3.3-70b-versatile")
+const openRouter = OpenRouter.configure({ apiKey: process.env.OPENROUTER_API_KEY ?? "fixture" })
+const openrouter = openRouter.model("openai/gpt-4o-mini")
+const openrouterGpt55 = openRouter.model("openai/gpt-5.5")
+const openrouterOpus = OpenRouter.configure({
+  apiKey: process.env.OPENROUTER_API_KEY ?? "fixture",
+}).model("anthropic/claude-opus-4.7")
+
+const redactCloudflareURL = (url: string) =>
+  url
+    .replace(/\/client\/v4\/accounts\/[^/]+\/ai\/v1\//, "/client/v4/accounts/{account}/ai/v1/")
+    .replace(/\/v1\/[^/]+\/[^/]+\/compat\//, "/v1/{account}/{gateway}/compat/")
+
+const cloudflareOptions = {
+  redactor: Redactor.defaults({ url: { transform: redactCloudflareURL } }),
+}
+
+describeRecordedGoldenScenarios([
+  {
+    name: "OpenAI Chat gpt-4o-mini",
+    prefix: "openai-chat",
+    model: openAIChat,
+    requires: ["OPENAI_API_KEY"],
+    scenarios: ["text", "tool-call", "tool-loop"],
+  },
+  {
+    name: "OpenAI Responses gpt-5.5",
+    prefix: "openai-responses",
+    model: openAIResponses,
+    requires: ["OPENAI_API_KEY"],
+    tags: ["flagship"],
+    scenarios: [
+      { id: "text", temperature: false },
+      { id: "reasoning", temperature: false },
+      { id: "reasoning-continuation", temperature: false },
+      { id: "tool-call", temperature: false },
+      { id: "tool-loop", temperature: false },
+      { id: "image-tool-result", temperature: false, maxTokens: 40 },
+    ],
+  },
+  {
+    name: "OpenAI Responses WebSocket gpt-4.1-mini",
+    prefix: "openai-responses-websocket",
+    model: openAIResponsesWebSocket,
+    transport: "websocket",
+    requires: ["OPENAI_API_KEY"],
+    scenarios: ["tool-loop"],
+  },
+  {
+    name: "Anthropic Haiku 4.5",
+    prefix: "anthropic-messages",
+    model: anthropicHaiku,
+    requires: ["ANTHROPIC_API_KEY"],
+    options: { redactor: Redactor.defaults({ requestHeaders: { allow: ["content-type", "anthropic-version"] } }) },
+    scenarios: ["text", "tool-call"],
+  },
+  {
+    name: "Anthropic Opus 4.7",
+    prefix: "anthropic-messages",
+    model: anthropicOpus,
+    requires: ["ANTHROPIC_API_KEY"],
+    tags: ["flagship"],
+    options: { redactor: Redactor.defaults({ requestHeaders: { allow: ["content-type", "anthropic-version"] } }) },
+    scenarios: [
+      { id: "tool-loop", temperature: false },
+      { id: "image-tool-result", temperature: false, maxTokens: 40 },
+    ],
+  },
+  {
+    name: "Gemini 2.5 Flash",
+    prefix: "gemini",
+    model: gemini,
+    requires: ["GOOGLE_GENERATIVE_AI_API_KEY"],
+    scenarios: [{ id: "text", maxTokens: 80 }, "tool-call", { id: "image", maxTokens: 160 }],
+  },
+  {
+    name: "xAI Grok 3 Mini",
+    prefix: "xai",
+    model: xaiBasic,
+    requires: ["XAI_API_KEY"],
+    scenarios: ["text", "tool-call"],
+  },
+  {
+    name: "xAI Grok 4.3",
+    prefix: "xai",
+    model: xaiFlagship,
+    requires: ["XAI_API_KEY"],
+    tags: ["flagship"],
+    scenarios: [{ id: "tool-loop", timeout: 30_000 }],
+  },
+  {
+    name: "Cloudflare AI Gateway Workers AI Llama 3.1 8B",
+    prefix: "cloudflare-ai-gateway",
+    model: cloudflareAIGatewayWorkers,
+    requires: ["CLOUDFLARE_ACCOUNT_ID", "CLOUDFLARE_API_TOKEN"],
+    options: cloudflareOptions,
+    scenarios: ["text"],
+  },
+  {
+    name: "Cloudflare AI Gateway Workers AI GPT OSS 20B Tools",
+    prefix: "cloudflare-ai-gateway",
+    model: cloudflareAIGatewayWorkersTools,
+    requires: ["CLOUDFLARE_ACCOUNT_ID", "CLOUDFLARE_API_TOKEN"],
+    options: cloudflareOptions,
+    scenarios: [{ id: "tool-call", maxTokens: 120 }],
+  },
+  {
+    name: "Cloudflare Workers AI Llama 3.1 8B",
+    prefix: "cloudflare-workers-ai",
+    model: cloudflareWorkersAI,
+    requires: ["CLOUDFLARE_ACCOUNT_ID", "CLOUDFLARE_API_KEY"],
+    options: cloudflareOptions,
+    scenarios: ["text"],
+  },
+  {
+    name: "Cloudflare Workers AI GPT OSS 20B Tools",
+    prefix: "cloudflare-workers-ai",
+    model: cloudflareWorkersAITools,
+    requires: ["CLOUDFLARE_ACCOUNT_ID", "CLOUDFLARE_API_KEY"],
+    options: cloudflareOptions,
+    scenarios: [{ id: "tool-call", maxTokens: 120 }],
+  },
+  {
+    name: "DeepSeek Chat",
+    prefix: "openai-compatible-chat",
+    model: deepseek,
+    requires: ["DEEPSEEK_API_KEY"],
+    scenarios: ["text"],
+  },
+  {
+    name: "TogetherAI Llama 3.3 70B",
+    prefix: "openai-compatible-chat",
+    model: together,
+    requires: ["TOGETHER_AI_API_KEY"],
+    scenarios: ["text", "tool-call"],
+  },
+  {
+    name: "Groq Llama 3.3 70B",
+    prefix: "openai-compatible-chat",
+    model: groq,
+    requires: ["GROQ_API_KEY"],
+    scenarios: ["text", "tool-call", { id: "tool-loop", timeout: 30_000 }],
+  },
+  {
+    name: "OpenRouter gpt-4o-mini",
+    prefix: "openai-compatible-chat",
+    model: openrouter,
+    requires: ["OPENROUTER_API_KEY"],
+    scenarios: ["text", "tool-call", "tool-loop"],
+  },
+  {
+    name: "OpenRouter gpt-5.5",
+    prefix: "openai-compatible-chat",
+    model: openrouterGpt55,
+    requires: ["OPENROUTER_API_KEY"],
+    tags: ["flagship"],
+    scenarios: ["tool-loop"],
+  },
+  {
+    name: "OpenRouter Claude Opus 4.7",
+    prefix: "openai-compatible-chat",
+    model: openrouterOpus,
+    requires: ["OPENROUTER_API_KEY"],
+    tags: ["flagship"],
+    scenarios: ["tool-loop"],
+  },
+])