Skip to content

Commit d34ecab

Browse files
authored
💄 style(openrouter): add stable versions of Gemini 2.5 models (lobehub#8239)
Signed-off-by: Max Brauer <mbrauer@vmware.com>
1 parent c7d4399 commit d34ecab

File tree

1 file changed

+36
-0
lines changed

1 file changed

+36
-0
lines changed

src/config/aiModels/openrouter.ts

Lines changed: 36 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -227,6 +227,24 @@ const openrouterChatModels: AIChatModelCard[] = [
227227
},
228228
type: 'chat',
229229
},
230+
{
231+
abilities: {
232+
functionCall: true,
233+
reasoning: true,
234+
vision: true,
235+
},
236+
contextWindowTokens: 1_048_576,
237+
description:
238+
'Gemini 2.5 Pro 是 Google 最先进的思维模型,能够对代码、数学和STEM领域的复杂问题进行推理,以及使用长上下文分析大型数据集、代码库和文档。',
239+
displayName: 'Gemini 2.5 Pro',
240+
id: 'google/gemini-2.5-pro',
241+
maxOutput: 65_536,
242+
pricing: {
243+
input: 1.25,
244+
output: 10,
245+
},
246+
type: 'chat',
247+
},
230248
{
231249
abilities: {
232250
functionCall: true,
@@ -245,6 +263,24 @@ const openrouterChatModels: AIChatModelCard[] = [
245263
},
246264
type: 'chat',
247265
},
266+
{
267+
abilities: {
268+
functionCall: true,
269+
reasoning: true,
270+
vision: true,
271+
},
272+
contextWindowTokens: 1_048_576,
273+
description:
274+
'Gemini 2.5 Flash 是 Google 最先进的主力模型,专为高级推理、编码、数学和科学任务而设计。它包含内置的“思考”能力,使其能够提供具有更高准确性和细致上下文处理的响应。\n\n注意:此模型有两个变体:思考和非思考。输出定价根据思考能力是否激活而有显著差异。如果您选择标准变体(不带“:thinking”后缀),模型将明确避免生成思考令牌。\n\n要利用思考能力并接收思考令牌,您必须选择“:thinking”变体,这将产生更高的思考输出定价。\n\n此外,Gemini 2.5 Flash 可通过“推理最大令牌数”参数进行配置,如文档中所述 (https://openrouter.ai/docs/use-cases/reasoning-tokens#max-tokens-for-reasoning)。',
275+
displayName: 'Gemini 2.5 Flash',
276+
id: 'google/gemini-2.5-flash',
277+
maxOutput: 65_535,
278+
pricing: {
279+
input: 0.15,
280+
output: 0.6,
281+
},
282+
type: 'chat',
283+
},
248284
{
249285
abilities: {
250286
functionCall: true,

0 commit comments

Comments
 (0)