DATA CUT 2026-07-27 116 活跃模型 680 A 类案例

CASE EVIDENCE / A RECORD

latitude-dev / latitude-llm 使用 GPT-5.1 (high) 处理知识检索和问答

latitude-dev / latitude-llm 公开的知识库与检索问答案例,来源为 公开代码库,复核于 2026-06-27T10:25:47Z。

原始记录:Latitude fixes Azure GPT-5.1 high-effort prompt routing for its LLM platform

A

Chinese Brief

中文案例导读

latitude-dev / latitude-llm 公开的知识库与检索问答案例,来源为 公开代码库,复核于 2026-06-27T10:25:47Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。 Model Atlas 不把 benchmark、教程、发布说明或集合页包装成真实案例。

厂商OpenAI
模型GPT-5.1 (high)
任务类型知识库与检索问答
审核状态auto_approved

任务

真实任务背景

这是一个围绕知识检索和问答的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:Latitude's LLM platform fixed Azure provider option routing so prompt configurations using model gpt-5.1 with reasoning_effort set to high are delivered under the OpenAI provider-options key expected by the Vercel AI SD…

知识库与检索问答公开代码库A 类可核验real_case
公开产物

公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:The patch adds regression coverage showing a Latitude config with model gpt-5.1 and reasoning_effort high is transformed into providerOptions.openai.reasoningEffort = high, preserving Azure resource metadata.

模型作用

GPT-5.1 (high) 在该案例中承担知识检索和问答相关的生成、分析、编排或实现角色。 原始资料写作:GPT-5.1 high supplies the reasoning-heavy execution mode for Latitude prompt runs on Azure OpenAI; the change ensures the model actually receives the high-effort instruction instead of silently dropping it.

风险边界

当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Evidence is a targeted bug-fix PR with explicit gpt-5.1 high configuration in tests; it verifies request routing rather than publishing user-facing generated outputs.