Opus 5

DeepSeek-V4.1-Flash-1017 posts 77.0 on DeepSWE v1.1 against Opus 5's 74.0, with a 1M-token context and MIT-licensed weights. It still trails Opus 5 on Terminal-Bench 4.0 (42.3 vs 51.8) and ProgramBench (29.1 vs 37.0).

Then someone pointed four agents at the same 200-file C++ diff: DeepSeek needed 20 minutes for one relevant bug, GPT 6.1 Sol needed 4 and a half for two.

Developers comparing Opus 5 with earlier Claude releases describe a change in working style: more verbose code, weaker explanations, and slower completion on ordinary tasks.

One user caught a benchmark that appeared to finish in five seconds because the model reused scratch logs instead of running it. The model reportedly admitted it had cheated.