# DeepSeek V4.1 Flash 1017 cuts reasoning tokens on ARC-AGI games

> Original title: @deepseek_ai @baseten At max reasoning with the provider adapter harness, DeepSeek V4.1 Flash 1017 used 36% fewer reasoning tokens on its...

- Source: ARC Prize
- Published: 2026-11-02T20:57:58.000Z
- Bad Influence: https://badinfluence.news/items/t8wmltmicgnyz0ul7nytfl8zw
- Original: https://x.com/arcprize/status/2117360592436527567

## Summary

ARC Prize says DeepSeek V4.1 Flash 1017, run on Baseten at max reasoning, used 36% fewer reasoning tokens on its first action than the original across 25 public games. On 55 semi-private games, the model used 48.6% fewer reasoning tokens on the first action.

## Post

@deepseek_ai @baseten At max reasoning with the provider adapter harness, DeepSeek V4.1 Flash 1017 used 36% fewer reasoning tokens on its first action than the original, averaged across 25 public games. Across 55 semi-private games, it used 48.6% fewer reasoning tokens on the first action.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWJgPATWgAAfXBu.jpg&mode=thumb&exp=1793923200&sig=368587a92db1b8e5)