OpenAI 前 CTO Mira Murati 创办的 Thinking Machines 发布新模型 Inkling 2 Small。Artificial Analysis…
This is a dev post classified by AGK Intelligence as AI dev tools (a launch), kept by the Dev Radar because it carries real work, not commentary.
OpenAI 前 CTO Mira Murati 创办的 Thinking Machines 发布新模型 Inkling 2 Small。Artificial Analysis 综合评分达到 31 分，比上一代 Inkling-Small 的 13 分提高了 18 分，成为美国得分最高的开源模型。 模型名字叫 Small，实际却有 5940 亿总参数，每次推理激活 170 亿。Agent 能力提升尤其明显：自动化任务通过率从 5% 涨到 46%，编程测试 FrontierCode 2.0 从 5% 涨到 25.4%，但仍低于 DeepSeek V4.1 Flash 的 28.5%。 Thinking Machines 对模型进行了持续 9 天的大规模强化学习训练，累计生成超过 1700 万条训练轨迹，覆盖编程、终端操作、文档制作、网页开发和工具调用等任务。 性能提升也带来了更高的消耗。Artificial Analysis 测试中，模型平均每项任务输出约 15 万 Token，是上一代的 4 倍。单项任务的平均评测成本从 0.20 美元涨到 1.18 美元，耗时约 21 分钟。 目前，Inkling 2 Small 已在 Tinker 开放推理和微调，但完整权重尚未发布。Thinking Machines 计划完成外部安全测试和红队测试后，下周以 Apache 2.0 许可
Posted by 思维怪怪 (8.5k followers) 59 min ago · 1 likes · 225 views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: Thinking Machines Lab.
More dev work like this
- Attention all Claude Code users ! — @ItsmeAjayKV
- 😱惊了啊！100万行神级DeepSeek Harness的框架，被他用1000行TypeScript彻底扒光！ — @NFT_Chen
- 想学 AI Agent，搜出来的教程东一篇西一篇，有的上来就讲多 Agent 协作，连 Token 是什么都还没弄明白。 — @GitHub_Daily
- Build an AI workflow visually, then publish it as a chat app. — @DanKornas
- I found a setup that saves you $300 a month on AI tools and you don't lose much — @VaibhavSisinty
- Hermes supremacy tweet #9138 — @Teknium
- A 9B model can help a much larger model get better at coding and improving itself. — @Chinazhidx
- Still need a monitor just to test your @NVIDIARobotics Jetson? — @seeedstudio
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.9k posts from 5k X accounts over the last 21 days, 9.5k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 07:57 UTC. Full method.