deepseek-v4.1VIDEO
DeepSeek V4.1 Flash 1017 runs Tool eval bench hard mode, scores coming
The author is running the Tool eval bench hard mode on DeepSeek V4.1-Flash-1017 in a sparkline dashboard and will share scores after multiple runs.
Sufyan
@sfxnzArchived 2 days agoOriginally posted 2026.10.27
Sufyan @sfxnzOriginally posted 2026.10.27
Tool eval bench hard mode running in the sparkline dashboard for my DeepSeek V4.1-Flash-1017. Sharing the scores soon after multiple runs.
x.com/sfxnz/status/2114828948026929566Archived 2026.10.27 12:09