GPT-6.1 Sol Cuts Mouse Brain Surgery Landmark Time from 169s to 86s via In-Context Learning
Exin Therapeutics CEO Gabriel Ocana-Santero reported that GPT-6.1 Sol participated in three real mouse brain surgeries, using in-context learning to improve performance. By the third operation, landmark localization time dropped from 169 to 86 seconds, corrections needed fell from 4 to 2, and positioning error decreased from 0.6mm to 0.3mm. The CEO confirmed the surgeries were real, not simulations, though the sample size is only three operations.
Coverage timeline
机器之心机器之心
在小鼠脑部手术中，考验操作者的往往是进针之前的一个细节：找到头骨上几个极其关键的解剖标志。这些点看起来不起眼，却决定了后续手术坐标如何计算。位置稍有偏差，针头就可能无法准确抵达预定的脑区。 现在，AI 也开始学习这项操作了。 近日，生物科技初创公司 Exin Therapeutics 联合创始人兼 CEO Gabriel Ocana-Santero 分享了一组实验结果：团队让 GPT-6.1 Sol 连续参与三轮小鼠脑部手术定位。 每次操作结束后，专业外科医生都会纠正模型的定位结果，模型则保存自己的图像、测量数据和专家修正记录，供下一轮操作参考。 结果相当有意思。 到了第三次操作， GPT-6.1 Sol 寻找解剖标志的时间从 169 秒缩短到 86 秒，所需的纠正动作从 4 次减少到 2 次，定位误差也从 0.6 毫米降至 0.3 毫米 。 Ocana-Santero 将这一过程称为基于真实世界经验的上下文学习（In-context Learning）。 需要注意的是，这组数据只有 3 次操作，距离证明 AI 已经掌握稳定的手术能力还有很长距离。但它展示了一种值得关注的变化：AI 开始把真实操作中的观察结果和专家反馈带入下一轮任务，逐步积累经验。 这项实验还引出了一个值得关注的问题：这究竟是真实手术，还是模拟环境中的演示？ X 用户 Stephane 在评论中直接向 Ocana-Santero 求证，后者明确回应是真实手术。Stephane 随后表示，这种结合上下文、在执行过程中即时积累经验的方式很有意思。他还透露，自己正在另一个领域探索类似机制：让模型保存每次任务中获得的经验，等再次遇到相关问题时，就能直接调用。 先找到头骨上的两个点， 才能准确定位大脑 要理解这次实验，得先认识两个神经科学实验中经常出现的名字：Bregma（前囟点）和 Lambda（人字缝点）。 它们并不是大脑内部的神经核团，而是小鼠头骨表面的两个解剖标志点。 Bregma 位于矢状缝与冠状缝的交汇处，Lambda 则位于矢状缝与人字缝的交汇处。研究人员通常以 Bregma 作为立体定位坐标的参考原点，再结合 Lambda 等标志校准头骨的位置和倾斜程度。 图源：https://www.jove.com/t/67381/an-immunocompetent-murine-model-for-l
Related stories
OpenAI releases GPT-6.1 Terra with computer-use focus, API price halved
OpenAI announced GPT-6.1 Terra, a fast, efficient model emphasizing visual reasoning and computer use, priced at $1 per million input tokens and $5 per million output tokens. The standard API price is 50% lower than GPT-5.6 Terra and GPT-6.1 Sol, and the model is available in ChatGPT Work, Codex, and via API (gpt-6.1-terra) for Plus, Pro, Enterprise, and Edu users. In the Agents' Last Exam benchmark, Terra scored 15 points higher than GPT-5.6 Terra at medium reasoning intensity, with about one-third the cost, and matched GPT-6 Astra's performance at one-tenth the token price.
4 reports · 4 sources
OpenAI, PerturbAI, Scripps deploy thousands of AI agents to analyze brain-cell CRISPR atlas in 32 minutes
On October 21, 2026, OpenAI, PerturbAI, and Scripps Research announced a study using thousands of AI agents powered by the GPT-Rosalind architecture to analyze a whole-mouse-brain CRISPR perturbation atlas. The distributed workflow completed in 32 minutes, a task estimated to take human researchers months to a year, and aggregated results into a structured reasoning atlas linking gene perturbations to cellular responses and literature.
2 reports · 2 sources159 points · 105 comments
Realtime-Bench: Frontier LLMs achieve only 44% success on real-time interactive tasks
Researchers from NVIDIA, UT-Austin, and other institutions released Realtime-Bench, a benchmark of 50 interactive game and web tasks that require real-time decision-making. Testing 12 frontier models, the best performer (GPT-6-Astra) achieved only 44% success, highlighting LLMs' weakness in low-latency tasks.
METR: Anthropic models improving twice as fast, AI use drives ~80% of acceleration
A METR report, based on an 11-week internal assessment at Anthropic, finds model capability progress roughly doubled since October 2025, with AI-assisted research driving about 80% of the acceleration. The report also reveals three unreleased internal models (Model 2, 3, 4), with the strongest exceeding Opus 5.5 by about three months of progress. Anthropic redacted much of the most informative evidence from the public version, citing confidentiality concerns.
6 reports · 5 sources47 points · 6 comments
News updates every hour. Subscribe in your RSS reader to get new stories automatically.
Subscribe