Thinking Machines Lab fine-tunes Qwen3-235B with expert annotations, surpassing frontier models on financial judgment tasks
Original title:Learning to Replicate Expert Judgment in Financial Tasks
AI summary
Thinking Machines Lab and Bridgewater AIA Labs released research in which Qwen3-235B was fine-tuned on data annotated by expert investors. Average accuracy on six financial information screening tasks rose from 78.2% for the best frontier model to 84.7%, while inference cost fell 13.8x.
Key Takeaways
- Thinking Machines Lab and Bridgewater AIA Labs released research in which Qwen3-235B was…
- Average accuracy on six financial information screening tasks rose from 78.
- 2% for the best frontier model to 84.
- 7%,while inference cost fell 13.
Quick Facts
- Published
- Updated
- Category
- Papers & Research
- AI score
- 63
原文给出了完整的训练配方和消融数据,读者可以看到专家标注加微调如何在特定任务上超过前沿模型并大幅降低成本。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Thinking Machines Lab fine-tunes Qwen3-235B with expert annotations, surpassing frontier models on financial judgment tasks. June 30, 2026. https://new.nat6.net/en/news/ibh1qalgvw1r18xn4vnwhx69q/