scriptoriumadmin/Roo-Code
1
0
Fork 0
mirror of https://github.com/RooVetGit/Roo-Code.git synced 2026-10-06 02:47:56 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Improve OpenRouter model fetching (#2922)

Browse source
This commit is contained in:
Chris Estreich 2025-04-24 12:29:36 -07:00 • committed by GitHub
parent ad4782b766
commit b75379bed3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
14 changed files with 362 additions and 126 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

97
package-lock.json generated
View file

@ -93,6 +93,7 @@
"knip": "^5.44.4",
"lint-staged": "^15.2.11",
"mkdirp": "^3.0.1",
"nock": "^14.0.4",
"npm-run-all": "^4.1.5",
"prettier": "^3.4.2",
"rimraf": "^6.0.1",
@ -6673,6 +6674,24 @@
"zod": "^3.24.1"
}
},
"node_modules/@mswjs/interceptors": {
"version": "0.38.6",
"resolved": "https://registry.npmjs.org/@mswjs/interceptors/-/interceptors-0.38.6.tgz",
"integrity": "sha512-qFlpmObPqeUs4u3oFYv/OM/xyX+pNa5TRAjqjvMhbGYlyMhzSrE5UfncL2rUcEeVfD9Gebgff73hPwqcOwJQNA==",
"dev": true,
"license": "MIT",
"dependencies": {
"@open-draft/deferred-promise": "^2.2.0",
"@open-draft/logger": "^0.3.0",
"@open-draft/until": "^2.0.0",
"is-node-process": "^1.2.0",
"outvariant": "^1.4.3",
"strict-event-emitter": "^0.5.1"
},
"engines": {
"node": ">=18"
}
},
"node_modules/@noble/ciphers": {
"version": "1.2.1",
"resolved": "https://registry.npmjs.org/@noble/ciphers/-/ciphers-1.2.1.tgz",
@ -6750,6 +6769,31 @@
"node": ">= 8"
}
},
"node_modules/@open-draft/deferred-promise": {
"version": "2.2.0",
"resolved": "https://registry.npmjs.org/@open-draft/deferred-promise/-/deferred-promise-2.2.0.tgz",
"integrity": "sha512-CecwLWx3rhxVQF6V4bAgPS5t+So2sTbPgAzafKkVizyi7tlwpcFpdFqq+wqF2OwNBmqFuu6tOyouTuxgpMfzmA==",
"dev": true,
"license": "MIT"
},
"node_modules/@open-draft/logger": {
"version": "0.3.0",
"resolved": "https://registry.npmjs.org/@open-draft/logger/-/logger-0.3.0.tgz",
"integrity": "sha512-X2g45fzhxH238HKO4xbSr7+wBS8Fvw6ixhTDuvLd5mqh6bJJCFAPwU9mPDxbcrRtfxv4u5IHCEH77BmxvXmmxQ==",
"dev": true,
"license": "MIT",
"dependencies": {
"is-node-process": "^1.2.0",
"outvariant": "^1.4.0"
}
},
"node_modules/@open-draft/until": {
"version": "2.1.0",
"resolved": "https://registry.npmjs.org/@open-draft/until/-/until-2.1.0.tgz",
"integrity": "sha512-U69T3ItWHvLwGg5eJ0n3I62nWuE6ilHlmz7zM0npLBRvPRd7e6NYmg54vvRtP5mZG7kZqZCFVdsTWo7BPtBujg==",
"dev": true,
"license": "MIT"
},
"node_modules/@pkgjs/parseargs": {
"version": "0.11.0",
"resolved": "https://registry.npmjs.org/@pkgjs/parseargs/-/parseargs-0.11.0.tgz",
@ -14204,6 +14248,13 @@
"url": "https://github.com/sponsors/ljharb"
}
},
"node_modules/is-node-process": {
"version": "1.2.0",
"resolved": "https://registry.npmjs.org/is-node-process/-/is-node-process-1.2.0.tgz",
"integrity": "sha512-Vg4o6/fqPxIjtxgUH5QLJhwZ7gW5diGCVlXpuUfELC62CuxM1iHcRe51f2W1FDy04Ai4KJkagKjx3XaqyfRKXw==",
"dev": true,
"license": "MIT"
},
"node_modules/is-number": {
"version": "7.0.0",
"resolved": "https://registry.npmjs.org/is-number/-/is-number-7.0.0.tgz",
@ -15525,6 +15576,13 @@
"integrity": "sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==",
"dev": true
},
"node_modules/json-stringify-safe": {
"version": "5.0.1",
"resolved": "https://registry.npmjs.org/json-stringify-safe/-/json-stringify-safe-5.0.1.tgz",
"integrity": "sha512-ZClg6AaYvamvYEE82d3Iyd3vSSIjQ+odgjaTzRuO3s7toCdFKczob2i0zCh7JE8kWn17yvAWhUVxvqGwUalsRA==",
"dev": true,
"license": "ISC"
},
"node_modules/json5": {
"version": "2.2.3",
"resolved": "https://registry.npmjs.org/json5/-/json5-2.2.3.tgz",
@ -16763,6 +16821,21 @@
"integrity": "sha512-1nh45deeb5olNY7eX82BkPO7SSxR5SSYJiPTrTdFUVYwAl8CKMA5N9PjTYkHiRjisVcxcQ1HXdLhx2qxxJzLNQ==",
"dev": true
},
"node_modules/nock": {
"version": "14.0.4",
"resolved": "https://registry.npmjs.org/nock/-/nock-14.0.4.tgz",
"integrity": "sha512-86fh+gIKH8H02+y0/HKAOZZXn6OwgzXvl6JYwfjvKkoKxUWz54wIIDU/+w24xzMvk/R8pNVXOrvTubyl+Ml6cg==",
"dev": true,
"license": "MIT",
"dependencies": {
"@mswjs/interceptors": "^0.38.5",
"json-stringify-safe": "^5.0.1",
"propagate": "^2.0.0"
},
"engines": {
"node": ">=18.20.0 <20 || >=20.12.1"
}
},
"node_modules/node-abi": {
"version": "3.74.0",
"resolved": "https://registry.npmjs.org/node-abi/-/node-abi-3.74.0.tgz",
@ -17428,6 +17501,13 @@
"integrity": "sha512-/jHxFIzoMXdqPzTaCpFzAAWhpkSjZPF4Vsn6jAfNpmbH/ymsmd7Qc6VE9BGn0L6YMj6uwpQLxCECpus4ukKS9Q==",
"dev": true
},
"node_modules/outvariant": {
"version": "1.4.3",
"resolved": "https://registry.npmjs.org/outvariant/-/outvariant-1.4.3.tgz",
"integrity": "sha512-+Sl2UErvtsoajRDKCE5/dBz4DIvHXQQnAxtQTF04OJxY0+DyZXSo5P5Bb7XYWOh81syohlYL24hbDwxedPUJCA==",
"dev": true,
"license": "MIT"
},
"node_modules/p-filter": {
"version": "2.1.0",
"resolved": "https://registry.npmjs.org/p-filter/-/p-filter-2.1.0.tgz",
@ -18120,6 +18200,16 @@
"node": ">= 6"
}
},
"node_modules/propagate": {
"version": "2.0.1",
"resolved": "https://registry.npmjs.org/propagate/-/propagate-2.0.1.tgz",
"integrity": "sha512-vGrhOavPSTz4QVNuBNdcNXePNdNMaO1xj9yBeH1ScQPjk/rhg9sSlCXPhMkFuaNNW/syTvYqsnbIJxMBfRbbag==",
"dev": true,
"license": "MIT",
"engines": {
"node": ">= 8"
}
},
"node_modules/proxy-addr": {
"version": "2.0.7",
"resolved": "https://registry.npmjs.org/proxy-addr/-/proxy-addr-2.0.7.tgz",
@ -19380,6 +19470,13 @@
"bare-events": "^2.2.0"
}
},
"node_modules/strict-event-emitter": {
"version": "0.5.1",
"resolved": "https://registry.npmjs.org/strict-event-emitter/-/strict-event-emitter-0.5.1.tgz",
"integrity": "sha512-vMgjE/GGEPEFnhFub6pa4FmJBRBVOLpIII2hvCZ8Kzb7K0hlHo7mQv6xYrBvCL2LtAIBwFUK8wvuJgTVSQ5MFQ==",
"dev": true,
"license": "MIT"
},
"node_modules/string_decoder": {
"version": "1.1.1",
"resolved": "https://registry.npmjs.org/string_decoder/-/string_decoder-1.1.1.tgz",

1
package.json
View file

@ -485,6 +485,7 @@
"knip": "^5.44.4",
"lint-staged": "^15.2.11",
"mkdirp": "^3.0.1",
"nock": "^14.0.4",
"npm-run-all": "^4.1.5",
"prettier": "^3.4.2",
"rimraf": "^6.0.1",

1
src/api/providers/__tests__/openai.test.ts
View file

@ -1,7 +1,6 @@
import { OpenAiHandler } from "../openai"
import { ApiHandlerOptions } from "../../../shared/api"
import { Anthropic } from "@anthropic-ai/sdk"
import { DEEP_SEEK_DEFAULT_TEMPERATURE } from "../constants"
// Mock OpenAI client
const mockCreate = jest.fn()

1
src/api/providers/__tests__/openrouter.test.ts
View file

@ -55,7 +55,6 @@ describe("OpenRouterHandler", () => {
info: mockOptions.openRouterModelInfo,
maxTokens: 1000,
reasoning: undefined,
supportsPromptCache: false,
temperature: 0,
thinking: undefined,
topP: undefined,

4
src/api/providers/constants.ts
View file

@ -6,7 +6,3 @@ export const DEFAULT_HEADERS = {
export const ANTHROPIC_DEFAULT_MAX_TOKENS = 8192
export const DEEP_SEEK_DEFAULT_TEMPERATURE = 0.6
export const AZURE_AI_INFERENCE_PATH = "/models/chat/completions"
export const REASONING_MODELS = new Set(["x-ai/grok-3-mini-beta", "grok-3-mini-beta", "grok-3-mini-fast-beta"])

25
src/api/providers/fetchers/__tests__/fixtures/openrouter-models.json Normal file
View file

File diff suppressed because one or more lines are too long

72
src/api/providers/fetchers/__tests__/openrouter.test.ts Normal file
View file

@ -0,0 +1,72 @@
// npx jest src/api/providers/fetchers/__tests__/openrouter.test.ts
import path from "path"
import { back as nockBack } from "nock"
import { PROMPT_CACHING_MODELS } from "../../../../shared/api"
import { getOpenRouterModels } from "../openrouter"
nockBack.fixtures = path.join(__dirname, "fixtures")
nockBack.setMode("dryrun")
describe("OpenRouter API", () => {
describe("getOpenRouterModels", () => {
it("fetches models and validates schema", async () => {
const { nockDone } = await nockBack("openrouter-models.json")
const models = await getOpenRouterModels()
expect(
Object.entries(models)
.filter(([_, model]) => model.supportsPromptCache)
.map(([id, _]) => id)
.sort(),
).toEqual(Array.from(PROMPT_CACHING_MODELS).sort())
expect(
Object.entries(models)
.filter(([_, model]) => model.supportsComputerUse)
.map(([id, _]) => id)
.sort(),
).toEqual([
"anthropic/claude-3.5-sonnet",
"anthropic/claude-3.5-sonnet:beta",
"anthropic/claude-3.7-sonnet",
"anthropic/claude-3.7-sonnet:beta",
"anthropic/claude-3.7-sonnet:thinking",
])
expect(models["anthropic/claude-3.7-sonnet"]).toEqual({
maxTokens: 8192,
contextWindow: 200000,
supportsImages: true,
supportsPromptCache: true,
inputPrice: 3,
outputPrice: 15,
cacheWritesPrice: 3.75,
cacheReadsPrice: 0.3,
description: expect.any(String),
thinking: false,
supportsComputerUse: true,
})
expect(models["anthropic/claude-3.7-sonnet:thinking"]).toEqual({
maxTokens: 128000,
contextWindow: 200000,
supportsImages: true,
supportsPromptCache: true,
inputPrice: 3,
outputPrice: 15,
cacheWritesPrice: 3.75,
cacheReadsPrice: 0.3,
description: expect.any(String),
thinking: true,
supportsComputerUse: true,
})
nockDone()
})
})
})

115
src/api/providers/fetchers/openrouter.ts Normal file
View file

@ -0,0 +1,115 @@
import axios from "axios"
import { z } from "zod"
import { ApiHandlerOptions, ModelInfo } from "../../../shared/api"
import { parseApiPrice } from "../../../utils/cost"
// https://openrouter.ai/api/v1/models
export const openRouterModelSchema = z.object({
id: z.string(),
name: z.string(),
description: z.string().optional(),
context_length: z.number(),
max_completion_tokens: z.number().nullish(),
architecture: z
.object({
modality: z.string().nullish(),
tokenizer: z.string().nullish(),
})
.optional(),
pricing: z
.object({
prompt: z.string().nullish(),
completion: z.string().nullish(),
input_cache_write: z.string().nullish(),
input_cache_read: z.string().nullish(),
})
.optional(),
top_provider: z
.object({
max_completion_tokens: z.number().nullish(),
})
.optional(),
})
export type OpenRouterModel = z.infer<typeof openRouterModelSchema>
const openRouterModelsResponseSchema = z.object({
data: z.array(openRouterModelSchema),
})
type OpenRouterModelsResponse = z.infer<typeof openRouterModelsResponseSchema>
export async function getOpenRouterModels(options?: ApiHandlerOptions) {
const models: Record<string, ModelInfo> = {}
const baseURL = options?.openRouterBaseUrl || "https://openrouter.ai/api/v1"
try {
const response = await axios.get<OpenRouterModelsResponse>(`${baseURL}/models`)
const result = openRouterModelsResponseSchema.safeParse(response.data)
const rawModels = result.success ? result.data.data : response.data.data
if (!result.success) {
console.error("OpenRouter models response is invalid", result.error.format())
}
for (const rawModel of rawModels) {
const cacheWritesPrice = rawModel.pricing?.input_cache_write
? parseApiPrice(rawModel.pricing?.input_cache_write)
: undefined
const cacheReadsPrice = rawModel.pricing?.input_cache_read
? parseApiPrice(rawModel.pricing?.input_cache_read)
: undefined
// Disable prompt caching for Gemini models for now.
const supportsPromptCache = !!cacheWritesPrice && !!cacheReadsPrice && !rawModel.id.startsWith("google")
const modelInfo: ModelInfo = {
maxTokens: rawModel.top_provider?.max_completion_tokens,
contextWindow: rawModel.context_length,
supportsImages: rawModel.architecture?.modality?.includes("image"),
supportsPromptCache,
inputPrice: parseApiPrice(rawModel.pricing?.prompt),
outputPrice: parseApiPrice(rawModel.pricing?.completion),
cacheWritesPrice,
cacheReadsPrice,
description: rawModel.description,
thinking: rawModel.id === "anthropic/claude-3.7-sonnet:thinking",
}
// Then OpenRouter model definition doesn't give us any hints about computer use,
// so we need to set that manually.
// The ideal `maxTokens` values are model dependent, but we should probably DRY
// this up and use the values defined for the Anthropic providers.
switch (true) {
case rawModel.id.startsWith("anthropic/claude-3.7-sonnet"):
modelInfo.supportsComputerUse = true
modelInfo.maxTokens = rawModel.id === "anthropic/claude-3.7-sonnet:thinking" ? 128_000 : 8192
break
case rawModel.id.startsWith("anthropic/claude-3.5-sonnet-20240620"):
modelInfo.maxTokens = 8192
break
case rawModel.id.startsWith("anthropic/claude-3.5-sonnet"):
modelInfo.supportsComputerUse = true
modelInfo.maxTokens = 8192
break
case rawModel.id.startsWith("anthropic/claude-3-5-haiku"):
case rawModel.id.startsWith("anthropic/claude-3-opus"):
case rawModel.id.startsWith("anthropic/claude-3-haiku"):
modelInfo.maxTokens = 8192
break
default:
break
}
models[rawModel.id] = modelInfo
}
} catch (error) {
console.error(
`Error fetching OpenRouter models: ${JSON.stringify(error, Object.getOwnPropertyNames(error), 2)}`,
)
}
return models
}

4
src/api/providers/openai.ts
View file

@ -15,7 +15,9 @@ import { convertToSimpleMessages } from "../transform/simple-format"
import { ApiStream, ApiStreamUsageChunk } from "../transform/stream"
import { BaseProvider } from "./base-provider"
import { XmlMatcher } from "../../utils/xml-matcher"
import { DEEP_SEEK_DEFAULT_TEMPERATURE, DEFAULT_HEADERS, AZURE_AI_INFERENCE_PATH } from "./constants"
import { DEFAULT_HEADERS, DEEP_SEEK_DEFAULT_TEMPERATURE } from "./constants"
export const AZURE_AI_INFERENCE_PATH = "/models/chat/completions"
export interface OpenAiHandlerOptions extends ApiHandlerOptions {}

119
src/api/providers/openrouter.ts
View file

@ -1,16 +1,19 @@
import { Anthropic } from "@anthropic-ai/sdk"
import { BetaThinkingConfigParam } from "@anthropic-ai/sdk/resources/beta"
import axios from "axios"
import OpenAI from "openai"
import { ApiHandlerOptions, ModelInfo, openRouterDefaultModelId, openRouterDefaultModelInfo } from "../../shared/api"
import { parseApiPrice } from "../../utils/cost"
import {
ApiHandlerOptions,
openRouterDefaultModelId,
openRouterDefaultModelInfo,
PROMPT_CACHING_MODELS,
} from "../../shared/api"
import { convertToOpenAiMessages } from "../transform/openai-format"
import { ApiStreamChunk } from "../transform/stream"
import { convertToR1Format } from "../transform/r1-format"
import { getModelParams, SingleCompletionHandler } from "../index"
import { DEFAULT_HEADERS, DEEP_SEEK_DEFAULT_TEMPERATURE } from "./constants"
import { getModelParams, SingleCompletionHandler } from ".."
import { BaseProvider } from "./base-provider"
const OPENROUTER_DEFAULT_PROVIDER_NAME = "[default]"
@ -62,15 +65,7 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
systemPrompt: string,
messages: Anthropic.Messages.MessageParam[],
): AsyncGenerator<ApiStreamChunk> {
let {
id: modelId,
maxTokens,
thinking,
temperature,
supportsPromptCache,
topP,
reasoningEffort,
} = this.getModel()
let { id: modelId, maxTokens, thinking, temperature, topP, reasoningEffort, info } = this.getModel()
// Convert Anthropic messages to OpenAI format.
let openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
@ -86,8 +81,8 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
// Prompt caching: https://openrouter.ai/docs/prompt-caching
// Now with Gemini support: https://openrouter.ai/docs/features/prompt-caching
// Note that we don't check the `ModelInfo` object because it is cached
// in the settings for OpenRouter.
if (this.isPromptCacheSupported(modelId)) {
// in the settings for OpenRouter and the value could be stale.
if (PROMPT_CACHING_MODELS.has(modelId)) {
openAiMessages[0] = {
role: "system",
// @ts-ignore-next-line
@ -191,7 +186,6 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
let id = modelId ?? openRouterDefaultModelId
const info = modelInfo ?? openRouterDefaultModelInfo
const supportsPromptCache = modelInfo?.supportsPromptCache
const isDeepSeekR1 = id.startsWith("deepseek/deepseek-r1") || modelId === "perplexity/sonar-reasoning"
const defaultTemperature = isDeepSeekR1 ? DEEP_SEEK_DEFAULT_TEMPERATURE : 0
const topP = isDeepSeekR1 ? 0.95 : undefined
@ -200,7 +194,6 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
id,
info,
...getModelParams({ options: this.options, model: info, defaultTemperature }),
supportsPromptCache,
topP,
}
}
@ -227,96 +220,4 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
const completion = response as OpenAI.Chat.ChatCompletion
return completion.choices[0]?.message?.content || ""
}
private isPromptCacheSupported(modelId: string) {
return (
modelId.startsWith("anthropic/claude-3.7-sonnet") ||
modelId.startsWith("anthropic/claude-3.5-sonnet") ||
modelId.startsWith("anthropic/claude-3-opus") ||
modelId.startsWith("anthropic/claude-3-haiku")
)
}
}
export async function getOpenRouterModels(options?: ApiHandlerOptions) {
const models: Record<string, ModelInfo> = {}
const baseURL = options?.openRouterBaseUrl || "https://openrouter.ai/api/v1"
try {
const response = await axios.get(`${baseURL}/models`)
const rawModels = response.data.data
for (const rawModel of rawModels) {
const modelInfo: ModelInfo = {
maxTokens: rawModel.top_provider?.max_completion_tokens,
contextWindow: rawModel.context_length,
supportsImages: rawModel.architecture?.modality?.includes("image"),
supportsPromptCache: false,
inputPrice: parseApiPrice(rawModel.pricing?.prompt),
outputPrice: parseApiPrice(rawModel.pricing?.completion),
description: rawModel.description,
thinking: rawModel.id === "anthropic/claude-3.7-sonnet:thinking",
}
// NOTE: This needs to be synced with api.ts/openrouter default model info.
switch (true) {
case rawModel.id.startsWith("anthropic/claude-3.7-sonnet"):
modelInfo.supportsComputerUse = true
modelInfo.supportsPromptCache = true
modelInfo.cacheWritesPrice = 3.75
modelInfo.cacheReadsPrice = 0.3
modelInfo.maxTokens = rawModel.id === "anthropic/claude-3.7-sonnet:thinking" ? 128_000 : 8192
break
case rawModel.id.startsWith("anthropic/claude-3.5-sonnet-20240620"):
modelInfo.supportsPromptCache = true
modelInfo.cacheWritesPrice = 3.75
modelInfo.cacheReadsPrice = 0.3
modelInfo.maxTokens = 8192
break
case rawModel.id.startsWith("anthropic/claude-3.5-sonnet"):
modelInfo.supportsComputerUse = true
modelInfo.supportsPromptCache = true
modelInfo.cacheWritesPrice = 3.75
modelInfo.cacheReadsPrice = 0.3
modelInfo.maxTokens = 8192
break
case rawModel.id.startsWith("anthropic/claude-3-5-haiku"):
modelInfo.supportsPromptCache = true
modelInfo.cacheWritesPrice = 1.25
modelInfo.cacheReadsPrice = 0.1
modelInfo.maxTokens = 8192
break
case rawModel.id.startsWith("anthropic/claude-3-opus"):
modelInfo.supportsPromptCache = true
modelInfo.cacheWritesPrice = 18.75
modelInfo.cacheReadsPrice = 1.5
modelInfo.maxTokens = 8192
break
case rawModel.id.startsWith("anthropic/claude-3-haiku"):
modelInfo.supportsPromptCache = true
modelInfo.cacheWritesPrice = 0.3
modelInfo.cacheReadsPrice = 0.03
modelInfo.maxTokens = 8192
break
/* TODO: uncomment once we confirm it's working
case rawModel.id.startsWith("google/gemini-2.5-pro-preview-03-25"):
case rawModel.id.startsWith("google/gemini-2.0-flash-001"):
case rawModel.id.startsWith("google/gemini-flash-1.5"):
modelInfo.supportsPromptCache = true
break
*/
default:
break
}
models[rawModel.id] = modelInfo
}
} catch (error) {
console.error(
`Error fetching OpenRouter models: ${JSON.stringify(error, Object.getOwnPropertyNames(error), 2)}`,
)
}
return models
}

8
src/api/providers/xai.ts
View file

@ -1,11 +1,13 @@
import { Anthropic } from "@anthropic-ai/sdk"
import OpenAI from "openai"
import { ApiHandlerOptions, XAIModelId, ModelInfo, xaiDefaultModelId, xaiModels } from "../../shared/api"
import { ApiHandlerOptions, XAIModelId, xaiDefaultModelId, xaiModels, REASONING_MODELS } from "../../shared/api"
import { ApiStream } from "../transform/stream"
import { convertToOpenAiMessages } from "../transform/openai-format"
import { DEFAULT_HEADERS, REASONING_MODELS } from "./constants"
import { SingleCompletionHandler } from "../index"
import { DEFAULT_HEADERS } from "./constants"
import { BaseProvider } from "./base-provider"
import { SingleCompletionHandler } from ".."
const XAI_DEFAULT_TEMPERATURE = 0

6
src/core/webview/webviewMessageHandler.ts
View file

@ -12,7 +12,7 @@ import { GlobalFileNames } from "../../shared/globalFileNames"
import { checkoutDiffPayloadSchema, checkoutRestorePayloadSchema, WebviewMessage } from "../../shared/WebviewMessage"
import { checkExistKey } from "../../shared/checkExistApiConfig"
import { EXPERIMENT_IDS, experimentDefault, ExperimentId } from "../../shared/experiments"
import { experimentDefault } from "../../shared/experiments"
import { Terminal } from "../../integrations/terminal/Terminal"
import { openFile, openImage } from "../../integrations/misc/open-file"
import { selectImages } from "../../integrations/misc/process-images"
@ -25,7 +25,7 @@ import { playTts, setTtsEnabled, setTtsSpeed, stopTts } from "../../utils/tts"
import { singleCompletionHandler } from "../../utils/single-completion-handler"
import { searchCommits } from "../../utils/git"
import { exportSettings, importSettings } from "../config/importExport"
import { getOpenRouterModels } from "../../api/providers/openrouter"
import { getOpenRouterModels } from "../../api/providers/fetchers/openrouter"
import { getGlamaModels } from "../../api/providers/glama"
import { getUnboundModels } from "../../api/providers/unbound"
import { getRequestyModels } from "../../api/providers/requesty"
@ -85,6 +85,7 @@ export const webviewMessageHandler = async (provider: ClineProvider, message: We
// if we hadn't retrieved the latest at this point
// (see normalizeApiConfiguration > openrouter).
const { apiConfiguration: currentApiConfig } = await provider.getState()
getOpenRouterModels(currentApiConfig).then(async (openRouterModels) => {
if (Object.keys(openRouterModels).length > 0) {
await provider.writeModelsToCache(GlobalFileNames.openRouterModels, openRouterModels)
@ -101,6 +102,7 @@ export const webviewMessageHandler = async (provider: ClineProvider, message: We
"openRouterModelInfo",
openRouterModels[apiConfiguration.openRouterModelId],
)
await provider.postStateToWebview()
}
}

32
src/shared/api.ts
View file

@ -1,7 +1,4 @@
import { ModelInfo, ProviderName, ProviderSettings } from "../schemas"
import { REASONING_MODELS } from "../api/providers/constants"
export { REASONING_MODELS }
export type { ModelInfo, ProviderName as ApiProvider }
@ -1399,3 +1396,32 @@ export const vscodeLlmModels = {
maxInputTokens: number
}
>
/**
* Constants
*/
export const REASONING_MODELS = new Set(["x-ai/grok-3-mini-beta", "grok-3-mini-beta", "grok-3-mini-fast-beta"])
export const PROMPT_CACHING_MODELS = new Set([
"anthropic/claude-3-haiku",
"anthropic/claude-3-haiku:beta",
"anthropic/claude-3-opus",
"anthropic/claude-3-opus:beta",
"anthropic/claude-3-sonnet",
"anthropic/claude-3-sonnet:beta",
"anthropic/claude-3.5-haiku",
"anthropic/claude-3.5-haiku-20241022",
"anthropic/claude-3.5-haiku-20241022:beta",
"anthropic/claude-3.5-haiku:beta",
"anthropic/claude-3.5-sonnet",
"anthropic/claude-3.5-sonnet-20240620",
"anthropic/claude-3.5-sonnet-20240620:beta",
"anthropic/claude-3.5-sonnet:beta",
"anthropic/claude-3.7-sonnet",
"anthropic/claude-3.7-sonnet:beta",
"anthropic/claude-3.7-sonnet:thinking",
// "google/gemini-2.0-flash-001",
// "google/gemini-flash-1.5",
// "google/gemini-flash-1.5-8b",
])

3
webview-ui/src/components/settings/constants.ts
View file

@ -9,10 +9,9 @@ import {
openAiNativeModels,
vertexModels,
xaiModels,
REASONING_MODELS,
} from "@roo/shared/api"
export { REASONING_MODELS }
export { REASONING_MODELS, PROMPT_CACHING_MODELS } from "@roo/shared/api"
export const MODELS_BY_PROVIDER: Partial<Record<ApiProvider, Record<string, ModelInfo>>> = {
anthropic: anthropicModels,