我们推出 GPT-6.1 Sol，这是对 GPT-6 Sol 的升级，在智能体编程、计算机使用和专业工作方面，其智能水平几乎与 GPT-6 Astra 相当，而标准输入和输出 token 价格仅为 Astra 的五分之一。缓存输入成本仅为每百万 token 0.10 美元——比标准输入价格低 95%，比 GPT-6 Sol 的缓存输入价格低 50%——为开发者提供了更多空间来构建和运行能够在多次请求间复用上下文的强大智能体。
GPT-6.1 Sol 为重要的日常工作提供了能力与成本的新平衡。在从编写和调试代码到理解文档和执行多步骤业务工作流的复杂专业任务上，它相比 GPT-6 Sol 实现了大幅改进。在其中若干评估中，它以显著更低的成本接近 GPT-6 Astra 的性能。
在 DeepSWE v1.1 上——该评估衡量真实代码库中的复杂软件工程任务——GPT-6.1 Sol 以约五分之一的成本达到与 GPT-6 Astra 相当的水平，同时在更低的推理努力和成本下，以 6.4 个百分点的优势超越了 GPT-6 Sol 的最佳得分。
在 GDP.pdf 上——该评估衡量模型使用复杂 PDF 文档（包括表格、图表、示意图和细小印刷细节）回答专业问题的准确程度——GPT-6.1 Sol 在测试的推理设置下，得分高于带 fallbacks 的 Opus 5.5，且每任务成本不到其一半。它还在每任务成本约为五分之一的情况下，接近 GPT-6 Astra 的最先进性能。
在 AutomationBench 上——该评估衡量智能体是否正确完成多步骤业务工作流——GPT-6.1 Sol 在中等推理努力下得分比 Opus 5.5 高 2.2 个百分点，成本约为其三分之一。在相同设置下，该得分也比 GPT-6 Sol 高出 4.8 个百分点。
GPT-6.1 Sol 在需要与计算机应用程序交互的任务上也取得了实质性进展。在 OSWorld 2.0 的离线集上——该评估在要求严苛的计算机使用工作流上衡量智能体——GPT-6.1 Sol 在最大推理努力下以不到一半的成本比 GPT-6 Sol 高出七个百分点。在最大推理努力下，其得分与 Astra 相差 2.1 个百分点以内，而每任务成本约为其七分之一。
在 Terminal-Bench Science 0.1 上——该评估衡量包括数据分析、模拟和定理证明在内的科学工作流——GPT-6.1 Sol 在最大推理努力下得分是 GPT-6 Sol 的两倍多，而每任务成本不到其一半。在最大努力下，GPT-6.1 Sol 平均每任务成本为 5.47 美元，而 Opus 5.5 为 23.21 美元，Astra 为 23.80 美元，以比这两个模型低 75% 以上的成本提供强大的科学能力。
GPT-6 Astra 在测试的模型中仍以 68.1% 取得最高得分，应将其用于最困难的科学研究任务。
GPT-6.1 Sol 还提高了在困难提示上的事实准确性。其相对于 GPT-6 Sol 的最大事实性改进出现在低推理努力下，将包含事实错误的回复占比从 11.4% 降至 7.7%——降低了约 32%。在测试的推理设置中，其错误率与 GPT-6 Astra 保持在 1.9 个百分点以内，而每任务成本不到其五分之一。
该评估衡量的是在用户标记了早期模型错误的去标识化对话中，包含至少一个事实错误的答案占比。这些刻意设置的困难提示并不代表典型使用情况。
GPT-6.1 Sol 在我们的对齐评估中相比 GPT-6 Sol 显示出实质性改进，使其更接近 GPT-6 Astra。
GPT-6.1 Sol 对其局限性更加透明，并且在尊重用户意图和安全约束方面更加可靠。在具有挑战性的评估中，它在搜索工具故障的透明度、遵守明确限制以及在智能体任务中避免未经授权的后果方面，失败率低于 GPT-6 Sol。我们未观察到任何绕过自动安全审查器的尝试，与 GPT-6 Astra 和 GPT-6 Sol 一致。完整细节可在 GPT-6.1 Sol 系统卡附录（在新窗口中打开）中找到。
以下评估刻意测试具有挑战性的情况，并不衡量典型使用中的失败率。
定价与可用性
GPT-6.1 Sol 从今天起面向 ChatGPT Work 和 Codex 中的所有 Plus、Pro、Business、Enterprise 和 Edu 用户提供。GPT-6.1 Sol 尚未在 Chat 中提供。开发者还可以通过 OpenAI API 以 gpt-6.1-sol 访问它。其标准 API 价格为每百万输入 token 2 美元、每百万缓存输入 token 0.10 美元、每百万输出 token 10 美元。未来几天，我们还将提供 GPT-6.1 Sol Ultrafast，在 Codex 中其 token 生成速度相比标准速度最高可快 8 倍。
We’re introducing GPT-6.1 Sol, an upgrade to GPT-6 Sol that nearly matches GPT-6 Astra’s intelligence on agentic coding, computer use, and professional work at one-fifth of Astra’s standard input and output token prices. Cached input costs just $0.10 per million tokens—95% less than standard input pricing and 50% less than GPT-6 Sol’s cached input pricing—giving developers more room to build and run capable agents that reuse context across requests.
GPT-6.1 Sol offers a new balance of capability and cost for important everyday work. It delivers substantial improvements over GPT-6 Sol across complex professional tasks, from writing and debugging code to understanding documents and executing multi-step business workflows. On several of these evaluations, it approaches GPT-6 Astra’s performance at substantially lower cost.
On DeepSWE v1.1, which evaluates complex software-engineering tasks in real codebases, GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth of the cost, while eclipsing GPT-6 Sol’s best score by 6.4 percentage points at a lower reasoning effort and cost.
On GDP.pdf, which measures how accurately models answer professional questions using complex PDF documents, including tables, charts, diagrams, and fine-print details, GPT-6.1 Sol scores higher than Opus 5.5 with fallbacks at less than half the cost per task across the tested reasoning settings. It also approaches GPT-6 Astra’s state-of-the-art performance at roughly one-fifth the cost per task.
On AutomationBench, which measures whether agents correctly complete multi-step business workflows, GPT-6.1 Sol scores 2.2 percentage points above Opus 5.5 at medium reasoning effort, at roughly a third of the cost. That score is also up 4.8 percentage points from GPT-6 Sol at the same setting.
GPT-6.1 Sol also makes substantial progress on tasks that require interacting with computer applications. On OSWorld 2.0’s offline set, which evaluates agents on demanding computer-use workflows, GPT-6.1 Sol outperforms GPT-6 Sol by seven percentage points at maximum reasoning effort at less than half the cost. It comes within 2.1 percentage points of Astra’s score at maximum reasoning effort at roughly one-seventh the cost per task.
On Terminal-Bench Science 0.1, which evaluates scientific workflows including data analysis, simulation, and theorem proving, GPT-6.1 Sol more than doubles GPT-6 Sol’s score at maximum reasoning effort at less than half the cost per task. At maximum effort, GPT-6.1 Sol costs $5.47 per task on average, compared with $23.21 for Opus 5.5 and $23.80 for Astra, delivering substantial scientific capability at over 75% lower cost than either model.
GPT-6 Astra still achieves the highest score among the models tested at 68.1%, and should be used for the most difficult scientific research tasks.
GPT-6.1 Sol also improves factual accuracy on difficult prompts. Its largest factuality improvement over GPT-6 Sol comes at low reasoning effort, where it reduces the share of responses containing a factual error from 11.4% to 7.7%—a reduction of approximately 32%. Across the tested reasoning settings, its error rate remains within 1.9 percentage points of GPT-6 Astra’s, at less than one-fifth the cost per task.
This evaluation measures the share of answers containing at least one factual error on de-identified conversations where users flagged an earlier model’s error. These deliberately difficult prompts are not representative of typical usage.
GPT-6.1 Sol shows substantial improvements over GPT-6 Sol in our alignment evaluations, bringing it closer to GPT-6 Astra.
GPT-6.1 Sol is more transparent about its limitations and more reliable at respecting user intent and safety constraints. In challenging evaluations, it shows lower failure rates than GPT-6 Sol on transparency about broken search tools, respecting explicit restrictions, and avoiding unauthorized outcomes during agentic tasks. We observed no attempts to bypass an automated safety reviewer, matching GPT-6 Astra and GPT-6 Sol. Full details can be found in the GPT-6.1 Sol system card addendum(opens in a new window).
The evaluations below deliberately test challenging situations and do not measure failure rates in typical use.
Pricing and availability
GPT-6.1 Sol is available starting today to all Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex. GPT-6.1 Sol is not yet available in Chat. Developers can also access it through the OpenAI API as gpt-6.1-sol. Its standard API prices are $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens. In the coming days, we’ll also offer GPT-6.1 Sol Ultrafast, with up to 8x faster token generation compared to its standard speed in Codex.
本文内容采集自官方网站，排版和翻译可能与原页面存在差异。
阅读官方全文