GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional work, and multi-step business workflow automation, and approaches Astra-level results on these tasks at a much lower cost. Compared with GPT-6 Sol, it makes fewer factual errors and is more reliable at respecting explicit restrictions and user intent during agentic tasks.
提供商
不同公司托管同一模型。OpenRouter 根据您选择的路由模式将请求路由到其中一家公司:Balanced(价格 + 速度)、Nitro(速度最快)或 Floor(价格最低)。
| US$1.00 | US$5.00 | US$0.05 | -- | -- |
- 输入 US$1.00 / 输出 US$5.00-- tok/s-- 正常运行时间
定价
客户为此模型实际支付的平均价格,与提供商公布的价格并列展示。由于缓存和折扣,实际支付的价格往往远低于标价。
加权平均值
GPT-6.1 Sol (batch) — 价格历史
| US$0.00 | US$0.00 | 12.1% | 100.0% |
- 有效输入价格 US$0.00 / 实际输出 US$0.00token 占比100.0%
性能
吞吐量是模型生成内容的速度(每秒 token 数,越高越好)。延迟是完整的往返时间(越低越好)。TTFT 是首个 token 的响应时间,即您看到内容出现前需要等待的时间(越低越好)。
正常运行时间
正常运行时间是指过去 3 天中至少有一家提供商响应请求的时间百分比。可用性是指成功提供推理服务的时间百分比。OpenRouter 会持续监控,并在某家提供商返回错误时使用次优提供商。
正常运行时间数据尚不足,暂时无法显示。
当上游提供商发生错误时,如果您的请求筛选条件允许,我们可以通过路由到另一家健康的提供商来恢复服务。您可以通过以下接口以编程方式访问各提供商的正常运行时间数据: 端点 API。了解更多有关我们的 负载均衡和提供商路由选项。
基准测试
标准化评测得分,以及此模型针对共用基准测试提示词的实际输出。百分比越高越好,每项排名显示此模型在该来源所评测的模型中的位置。
| 来源 | 基准测试 | 得分 |
|---|---|---|
| Artificial Analysis | GPT-6.1 Sol (Xhigh) Intelligence Index | 52.5 |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) HLE | 52.6% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) AA-LCR | 79.7% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) GDPval-AA | 53.1% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) CritPt | 31.7% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) SciCode | 55.7% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) Terminal-Bench 4.0 | 54.0% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) AA-Omniscience Accuracy | 61.1% |
| Artificial Analysis | GPT-6.1 Sol (Xhigh) AA-Omniscience Non-Hallucination Rate | 52.0% |
| Artificial Analysis | GPT-6.1 Sol (Low) Intelligence Index | 34.8 |
| Artificial Analysis | GPT-6.1 Sol (Low) HLE | 47.4% |
| Artificial Analysis | GPT-6.1 Sol (Low) AA-LCR | 84.0% |
| Artificial Analysis | GPT-6.1 Sol (Low) GDPval-AA | 41.1% |
| Artificial Analysis | GPT-6.1 Sol (Low) CritPt | 24.9% |
| Artificial Analysis | GPT-6.1 Sol (Low) SciCode | 53.2% |
| Artificial Analysis | GPT-6.1 Sol (Low) Terminal-Bench 4.0 | 30.8% |
| Artificial Analysis | GPT-6.1 Sol (Low) AA-Omniscience Accuracy | 58.6% |
| Artificial Analysis | GPT-6.1 Sol (Low) AA-Omniscience Non-Hallucination Rate | 50.1% |
| Artificial Analysis | GPT-6.1 Sol (High) Intelligence Index | 50.3 |
| Artificial Analysis | GPT-6.1 Sol (High) HLE | 51.4% |
| Artificial Analysis | GPT-6.1 Sol (High) AA-LCR | 82.3% |
| Artificial Analysis | GPT-6.1 Sol (High) GDPval-AA | 51.1% |
| Artificial Analysis | GPT-6.1 Sol (High) CritPt | 30.0% |
| Artificial Analysis | GPT-6.1 Sol (High) SciCode | 55.8% |
| Artificial Analysis | GPT-6.1 Sol (High) Terminal-Bench 4.0 | 51.5% |
| Artificial Analysis | GPT-6.1 Sol (High) AA-Omniscience Accuracy | 60.4% |
| Artificial Analysis | GPT-6.1 Sol (High) AA-Omniscience Non-Hallucination Rate | 51.0% |
| Artificial Analysis | GPT-6.1 Sol (Max) Intelligence Index | 54.6 |
| Artificial Analysis | GPT-6.1 Sol (Max) HLE | 52.9% |
| Artificial Analysis | GPT-6.1 Sol (Max) AA-LCR | 83.0% |
| Artificial Analysis | GPT-6.1 Sol (Max) GDPval-AA | 55.4% |
| Artificial Analysis | GPT-6.1 Sol (Max) CritPt | 31.7% |
| Artificial Analysis | GPT-6.1 Sol (Max) SciCode | 54.2% |
| Artificial Analysis | GPT-6.1 Sol (Max) Terminal-Bench 4.0 | 56.1% |
| Artificial Analysis | GPT-6.1 Sol (Max) AA-Omniscience Accuracy | 61.8% |
| Artificial Analysis | GPT-6.1 Sol (Max) AA-Omniscience Non-Hallucination Rate | 48.1% |
| Artificial Analysis | GPT-6.1 Sol (Medium) Intelligence Index | 46.3 |
| Artificial Analysis | GPT-6.1 Sol (Medium) HLE | 49.9% |
| Artificial Analysis | GPT-6.1 Sol (Medium) AA-LCR | 83.3% |
| Artificial Analysis | GPT-6.1 Sol (Medium) GDPval-AA | 48.2% |
| Artificial Analysis | GPT-6.1 Sol (Medium) CritPt | 27.7% |
| Artificial Analysis | GPT-6.1 Sol (Medium) SciCode | 53.2% |
| Artificial Analysis | GPT-6.1 Sol (Medium) Terminal-Bench 4.0 | 48.0% |
| Artificial Analysis | GPT-6.1 Sol (Medium) AA-Omniscience Accuracy | 59.6% |
| Artificial Analysis | GPT-6.1 Sol (Medium) AA-Omniscience Non-Hallucination Rate | 51.9% |
| Design Arena | GPT-6.1 Sol Agents Arena Androidnative Elo | 1194 |
| Design Arena | GPT-6.1 Sol Agents Arena Blender3d Elo | 1268 |
| Design Arena | GPT-6.1 Sol Agents Arena Codetovideo Elo | 1199 |
| Design Arena | GPT-6.1 Sol Agents Arena Mobile Apps Elo | 1226 |
| Design Arena | GPT-6.1 Sol Agents Arena Pdf Elo | 1125 |
| Design Arena | GPT-6.1 Sol Agents Arena Spreadsheet Elo | 1209 |
| Design Arena | GPT-6.1 Sol Agents Arena Webapps Elo | 1279 |
| Design Arena | GPT-6.1 Sol Models Arena 3D Elo | 1426 |
| Design Arena | GPT-6.1 Sol Models Arena Code Categories Elo | 1361 |
| Design Arena | GPT-6.1 Sol Models Arena Data Visualization Elo | 1278 |
| Design Arena | GPT-6.1 Sol Models Arena Game Development Elo | 1405 |
| Design Arena | GPT-6.1 Sol Models Arena SVG Elo | 1344 |
| Design Arena | GPT-6.1 Sol Models Arena UI Component Elo | 1352 |
| Design Arena | GPT-6.1 Sol Models Arena Website Elo | 1315 |
| Vals AI | GPT-6.1 Sol Access to Justice Benchmark | 65.8% |
| Vals AI | GPT-6.1 Sol BioMysteryBench | 79.6% |
| Vals AI | GPT-6.1 Sol ClaimBench | 100.0% |
| Vals AI | GPT-6.1 Sol Code Migration v1.1 | 77.3% |
| Vals AI | GPT-6.1 Sol CyberBench v1.1 | 39.3% |
| Vals AI | GPT-6.1 Sol EMB v1.1 | 71.1% |
| Vals AI | GPT-6.1 Sol Finance Agent v2.1 | 61.8% |
| Vals AI | GPT-6.1 Sol Vals Finance Index | 55.8% |
| Vals AI | GPT-6.1 Sol Financial Coaching Bench | 68.5% |
| Vals AI | GPT-6.1 Sol HLAB-Vals-1.1 | 6.3% |
| Vals AI | GPT-6.1 Sol Immigration | 67.8% |
| Vals AI | GPT-6.1 Sol IOI | 96.9% |
| Vals AI | GPT-6.1 Sol Vals Legal Index | 68.7% |
| Vals AI | GPT-6.1 Sol Legal Research Bench | 38.5% |
| Vals AI | GPT-6.1 Sol MedCode | 48.8% |
| Vals AI | GPT-6.1 Sol MedScribe | 86.5% |
| Vals AI | GPT-6.1 Sol MysteryMechanism | 46.4% |
| Vals AI | GPT-6.1 Sol ProgramBench | 3.0% |
| Vals AI | GPT-6.1 Sol ProofBench v1.1 | 99.0% |
| Vals AI | GPT-6.1 Sol Public Benefits Bench v1.1 | 59.3% |
| Vals AI | GPT-6.1 Sol Robo-bench | 28.3% |
| Vals AI | GPT-6.1 Sol Vals RSI Index | 26.6% |
| Vals AI | GPT-6.1 Sol SAGE | 90.5% |
| Vals AI | GPT-6.1 Sol Vals Science Index | 59.6% |
| Vals AI | GPT-6.1 Sol SRE Bench | 50.8% |
| Vals AI | GPT-6.1 Sol SWE-bench Multimodal | 43.5% |
| Vals AI | GPT-6.1 Sol Tax Agent Bench | 21.8% |
| Vals AI | GPT-6.1 Sol Terminal-Bench 4.0 | 55.1% |
| Vals AI | GPT-6.1 Sol Terminal-Bench-Science | 52.9% |
| Vals AI | GPT-6.1 Sol Vals Index | 63.9% |
| Vals AI | GPT-6.1 Sol Vibe Code Bench v1.1 | 88.9% |
应用
向此模型发送最多流量的公开应用。这能很好地反映实际生产工作负载的情况,也能提示此模型最适合哪些用例。
使用此模型的应用尚无足够数据。
活动
此模型的 token 用量和请求流量随时间的变化。
API
可直接用于调用此模型的代码。OpenRouter 的 API 与 OpenAI 兼容——大多数 SDK 只需替换基础 URL 即可使用。切换模型时,唯一需要更改的是下方的模型标识符。
目前无法通过 EU 区域内路由使用。 区域内路由文档
文档
在 OpenRouter 上使用 OpenAI: GPT-6.1 Sol (batch) 构建应用的指南、实用教程、博客文章和工具。
常见问题
GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional work, and multi-step business workflow automation, and approaches Astra-level results on these tasks at a much lower cost.
OpenRouter 上的 GPT-6.1 Sol (batch) 文档包括 Reranking in RAG: How to Fix Bad Retrieval with a Reranker 博客文章和How to Benchmark LLMs 博客文章。快速入门介绍了 API 密钥和您的首次请求,API 参考列出了所有参数。
GPT-6.1 Sol (batch) 的费用为 US$1.00/百万输入 token 和 US$5.00/百万输出 token,Cache Read,价格为 US$0.05/M tokens、Cache Write,价格为 US$1.25/M tokens和Web Search,价格为 US$10.00/1K calls 单独计费。
GPT-6.1 Sol (batch) 的上下文窗口为 1,050,000 个 token。它最多支持 128,000 个补全 token。
是的。GPT-6.1 Sol (batch) 接受 tools 和 tool_choice,以进行函数调用。 它还支持通过 response_format 中的 JSON schema 实现结构化输出。
是的。OpenAI 会自动缓存包含 1024 个或更多 token 的提示词,而 GPT-6.1 Sol (batch) 会对每个已缓存的提示词收取缓存写入费用,无论缓存前缀是否被再次读取。要为某个请求关闭缓存,请在请求根级别将 prompt_cache_options.mode 设为 "explicit",且不要为任何内容块标记 prompt_cache_breakpoint。显式模式会禁用 OpenAI 管理的断点,因此,如果请求未设置您自己的断点,就没有可缓存的前缀,并按普通输入费率计费。请参阅提示词缓存指南。
GPT-6.1 Sol (batch) 接受 PDF 等文件、图像和文本 作为输入,并返回 文本。
在 OpenRouter 上,GPT-6.1 Sol (batch) 由 1 家提供商提供服务:OpenAI。
GPT-6.1 Terra Pro, GPT-6.1 Terra, GPT-6 Luna Decisions 及另外 58 个 是 OpenAI 的其他 文本 模型。
GPT-6.1 Sol (batch) 于 2026年9月29日 发布。