deepseek-v4.1IMAGE
Updated coding bench: DeepSeek 1017 Flash second only to Opus 5.5, ahead of GPT-6.1 Sol and Astra
The author shares an updated coding benchmark screenshot showing DeepSeek 1017 Flash in second place behind Opus 5.5 and ahead of GPT-6.1 Sol and Astra, noting it ran in the Codex harness and would do better in Pi or DSH.
gary IH fung
@garyfungArchived 9 days agoOriginally posted 2026.10.23
gary IH fung @garyfungOriginally posted 2026.10.23
deepseek 1017 flash is second only to opus 5.5, beating gpt 6.1 sol and astra. On the updated coding bench 🐋 flash beating astra. What. & it's handicapped by Codex harness instead of Pi or DSH which it does better
This post quotes another post: “Artificial Analysis Coding Agent Index v2.0: Claude Opus 5.5 leads at 61.7”
x.com/garyfung/status/2113472942420197626Archived 2026.10.23 12:06