OpenAI发布GPT-6.2 Astra与Sol,API降价20%
OpenAI Releases GPT-6.2 Astra and Sol With a 20% Cheaper Astra API
GPT-6.2系列发布且伴随显著降价,直接影响开发者成本结构与Agent部署策略,建议关注API调用的性价比变化。
OpenAI has released GPT-6.2 Astra and GPT-6.2 Sol, twp closed models aimed at professional and agentic work. The GPT-6.2 Astra model is the new frontier tier for hard, long-horizon tasks. GPT-6.2 Sol is the efficient everyday tier. For developers, the key point is cost per task. Astra’s token prices fall 20%, and OpenAI says both models finish tasks with fewer output tokens. Both are live in the OpenAI API as gpt-6.2-astra and gpt-6.2-sol.
OpenAI 发布了 GPT-6.2 Astra 和 GPT-6.2 Sol,这两款闭源模型旨在服务于专业场景和智能体工作。GPT-6.2 Astra 模型是面向高难度、长周期任务的前沿层级。GPT-6.2 Sol 则是高效日常层级。对于开发者而言,关键在于每项任务的成本。Astra 的 Token 价格下降了 20%,OpenAI 表示这两款模型都能以更少的输出 Token 完成任务。两者均已上线 OpenAI API,分别对应 gpt-6.2-astra 和 gpt-6.2-sol。
TL;DR
TL;DR(太长不看)
- Size: 1,050,000-token context window, 922,000 max input, 128,000 max output.
- Runs on: OpenAI’s hosted API, ChatGPT Work and Codex only.
- Performance: Astra scores 62 on the Artificial Analysis Intelligence Index, second behind Claude Fable 5.2 (65).
- Best: Astra ranks #1 on AA-3D with 1,318 Elo, ahead of Fable 5.2 (1,271).
- Worst: Astra posts a 59% AA-Omniscience hallucination rate (Sol: 55%). Sol’s Coding Agent Index drops 3 points.
- Bottom line: Best: frontier results at much lower cost per task. Worst: no coding gains, and hallucination stays high.
- 规模:1,050,000 Token 上下文窗口,最大输入 922,000 Token,最大输出 128,000 Token。
- 运行平台:仅限 OpenAI 托管 API、ChatGPT Work 和 Codex。
- 性能:在 Artificial Analysis Intelligence Index 中,Astra 得分为 62,仅次于 Claude Fable 5.2(65)。
- 最佳表现:在 AA-3D 基准测试中,Astra 以 1,318 Elo 排名 #1,领先于 Fable 5.2(1,271)。
- 最差表现:Astra 在 AA-Omniscience 幻觉率为 59%(Sol 为 55%)。Sol 的 Coding Agent Index 下降 3 分。
- 总结:最佳——以更低的单任务成本实现前沿级结果;最差——编码能力无提升,且幻觉率依然较高。
What is GPT-6.2 Astra?
什么是 GPT-6.2 Astra?
GPT-6.2 Astra is OpenAI’s frontier reasoning model for complex professional work. It succeeds GPT-6 Astra. OpenAI says it produces cleaner slides, spreadsheets and documents that follow templates and field conventions. It also builds better frontends, games and 3D scenes.
GPT-6.2 Astra 是 OpenAI 面向复杂专业工作的前沿推理模型,它是 GPT-6 Astra 的继任者。OpenAI 表示它能生成更整洁的幻灯片、电子表格和文档,遵循模板和字段规范。它还能构建更好的前端、游戏和 3D 场景。
GPT-6.2 Sol replaces GPT-6.1 Sol as the everyday model.
GPT-6.2 Sol 取代 GPT-6.1 Sol 成为日常使用模型。
How do the two models differ?
这两款模型有何区别?
Both accept text and image input and return text. Both share the same context limits. The split is cost and depth.
两者均接受文本和图片输入并返回文本。两者共享相同的上下文限制。主要差异在于成本和深度。
- Astra: knowledge cutoff May 18, 2026. Built for the hardest multistep work.
- Sol: knowledge cutoff June 1, 2026. Built for fast, versatile everyday tasks.
- Astra:知识截止日期为 2026 年 5 月 18 日。专为最复杂的多元步骤工作而设计。
- Sol:知识截止日期为 2026 年 6 月 1 日。专为快速、通用的日常任务而设计。
Neither model accepts reasoning effort none. OpenAI recommends keeping prompt prefixes unchanged and appending new instructions. That improves cache reuse and preserves reasoning across turns.
两款模型均不支持“无推理努力”模式。OpenAI 建议保持提示词前缀不变,并在其后追加新指令。这有助于提高缓存复用率,并在多轮对话中保持推理一致性。
What do the benchmarks show?
基准测试结果如何?
OpenAI’s own claims focus on cost-adjusted gains:
OpenAI 自身的声明侧重于成本调整后的收益:
- Agents’ Last Exam v1.1: Astra scores 5.2 points above GPT-6 Astra at under a third of the cost. It leads Claude Fable 5.2 by 5.5 points at 5% of the cost.
- Terminal-Bench-Science 0.1: Astra gains 9 points over GPT-6 Astra at under a third of the cost.
- GDPval v2 (internal): Astra gains 6.2 points with 22% fewer output tokens. Sol gains 5.3 points over GPT-6.1 Sol.
- FrontierCode 2.0: Sol lands within 0.5 points of Astra at about one-fifth the cost.
- AA-AnalystAgent: Astra answers 62.5% of questions correctly on all 5 attempts.
- Agents’ Last Exam v1.1:Astra 得分比 GPT-6 Astra 高出 5.2 分,而成本不到三分之一。其得分比 Claude Fable 5.2 高出 5.5 分,但成本仅为后者的 5%。
- Terminal-Bench-Science 0.1:Astra 得分比 GPT-6 Astra 高出 9 分,而成本不到三分之一。
- GDPval v2(内部数据):Astra 得分提升 6.2 分,且输出 Token 减少 22%。Sol 得分比 GPT-6.1 Sol 高出 5.3 分。
- FrontierCode 2.0:Sol 得分与 Astra 相差仅 0.5 分,而成本约为 Astra 的五分之一。
- AA-AnalystAgent:Astra 在所有 5 次尝试中正确回答了 62.5% 的问题。
Independent results from Artificial Analysis mostly hold up. Astra scores 62 at max effort, with a $4.00 cost per Intelligence Index task. Fable 5.2 scores 65, Gemini 4 Carbon 60, Claude Opus 5.5 59 and Sol 57. Sol matches GPT-6 Astra’s score at about 15% of its cost per task.
来自 Artificial Analysis 的独立结果基本成立。Astra 在最大努力下的得分为 62,每个 Intelligence Index 任务的成本为 4.00 美元。Fable 5.2 得分为 65,Gemini 4 Carbon 为 60,Claude Opus 5.5 为 59,Sol 为 57。Sol 以约为其成本 15% 的价格达到了与 GPT-6 Astra 相同的分数。
On Harvey LAB-AA, Astra passes 15.0% of legal assignments. The Astra for Law configuration passes 20.8%. Epoch’s Capabilities Index puts Astra at 174, tied with Fable 5.2.
在 Harvey LAB-AA 上,Astra 通过了 15.0% 的法律任务。Astra for Law 配置通过了 20.8%。Epoch 的 Capabilities Index 将 Astra 排在 174 分,与 Fable 5.2 并列。
How much does GPT-6.2 cost in the API?
GPT-6.2 在 API 中的价格是多少?
| Rate (per 1M tokens) | GPT-6.2 Astra | GPT-6.2 Sol |
|---|---|---|
| Input | $8.00 | $2.00 |
| Cached input | $0.40 | $0.10 |
| Output | $40.00 | $10.00 |
| 费率(每百万 token) | GPT-6.2 Astra | GPT-6.2 Sol |
|---|---|---|
| 输入 | $8.00 | $2.00 |
| 缓存输入 | $0.40 | $0.10 |
| 输出 | $40.00 | $10.00 |
Astra’s input and output prices are 20% below GPT-6 Astra. Cached input is 60% cheaper. Sol keeps GPT-6.1 Sol’s rates. Requests over 272K input tokens double input and cache rates, and output costs 1.5x. Batch is 50% of standard. Ultrafast is 3x standard through February 28, 2027, then 6x.
Astra 的输入和输出价格比 GPT-6 Astra 低 20%。缓存输入便宜 60%。Sol 保持了 GPT-6.1 Sol 的费率。超过 272K 输入 token 的请求会使输入和缓存费率翻倍,输出成本为 1.5 倍。批量处理为标准价格的 50%。Ultrafast 在 2027 年 2 月 28 日之前为标准价格的 3 倍,之后为 6 倍。
How does it compare with competitors?
它与竞争对手相比如何?
| Feature | GPT-6.2 Astra | GPT-6.2 Sol | Claude Fable 5.2 | Gemini 4 Carbon |
|---|---|---|---|---|
| Parameters | Not disclosed | Not disclosed | Not disclosed | Not disclosed |
| Context window | 1,050,000 | 1,050,000 | 1,000,000 | 1,048,576 |
| Max output | 128,000 | 128,000 | 128,000 | Not disclosed |
| Input / output price | $8 / $40 | $2 / $10 | $8 / $40 | $2 / $10 (until Nov 30), then $4 / $20 |
| Cached input | $0.40 | $0.10 | $0.20 | $0.10 (intro) |
| AA Intelligence Index | 62 | 57 | 65 | 60 |
| AA-3D Elo | 1,318 | 1,259 | 1,271 | Not disclosed |
| AA-Robotics v0.1 | 79 | 78 | 75 | Not disclosed |
| Weights | Closed | Closed | Closed | Closed |
| 功能 | GPT-6.2 Astra | GPT-6.2 Sol | Claude Fable 5.2 | Gemini 4 Carbon |
|---|---|---|---|---|
| 参数 | 未披露 | 未披露 | 未披露 | 未披露 |
| 上下文窗口 | 1,050,000 | 1,050,000 | 1,000,000 | 1,048,576 |
| 最大输出 | 128,000 | 128,000 | 128,000 | 未披露 |
| 输入/输出价格 | $8 / $40 | $2 / $10 | $8 / $40 | $2 / $10(直到 11 月 30 日),之后为 $4 / $20 |
| 缓存输入 | $0.40 | $0.10 | $0.20 | $0.10(介绍期) |
| AA Intelligence Index | 62 | 57 | 65 | 60 |
| AA-3D Elo | 1,318 | 1,259 | 1,271 | 未披露 |
| AA-Robotics v0.1 | 79 | 78 | 75 | 未披露 |
| 权重 | 闭源 | 闭源 | 闭源 | 闭源 |
Index, AA-3D and robotics figures for competitors come from Artificial Analysis as reported by The Decoder.
竞争对手的指数、AA-3D 和机器人数据来自 Artificial Analysis,由 The Decoder 报道。
Astra is also OpenAI’s first released model to meet the High threshold for AI self-improvement.
Astra 也是 OpenAI 发布的第一个达到 AI 自我改进高阈值的模型。
Key Takeaways
关键要点
- Astra drops to $8/$40 per 1M tokens, 20% below GPT-6 Astra.
- Astra ranks #2 on the AA Intelligence Index with 62 points.
- Sol nearly matches Astra on FrontierCode 2.0 at one-fifth the cost.
- Coding agent scores did not improve, and Sol slipped 3 points.
- Both models ship with a 1.05M-token context and 128K max output.
- Astra 的价格降至每百万 token $8/$40,比 GPT-6 Astra 低 20%。
- Astra 在 AA Intelligence Index 上排名第 2,得分为 62 分。
- Sol 在 FrontierCode 2.0 上几乎与 Astra 持平,但成本仅为五分之一。
- 编码智能体得分没有提高,Sol 下滑了 3 分。
- 两款模型均配备 1.05M token 的上下文窗口和 128K 的最大输出。
Check out the OpenAI announcement, the GPT-6.2 Astra model page and the GPT-6.2 Sol model page. You can check the deployment options and specs here through our GPULeaderboard.com chrome extension.
查看 OpenAI 的公告、GPT-6.2 Astra 模型页面和 GPT-6.2 Sol 模型页面。您可以通过我们的 GPULeaderboard.com Chrome 扩展程序在此处查看部署选项和规格。
Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
此外,欢迎在 Twitter 上关注我们,别忘了加入我们有 15 万+成员的 ML SubReddit 并订阅我们的新闻通讯。等等!你在 Telegram 上吗?现在你也可以加入我们的 Telegram 群组。
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力