deepseek-v4.1POST
DeepSeek V4.1 Flash 1017 vs Opus 5 and GPT-5.6 Sol: DeepSWE 77.0% at ~$0.2 per task
Official charts put it at 77.0% on DeepSWE v1.1, above Opus 5's 74.0 and GPT-5.6-Sol's 73.0, while still trailing Opus 5 on HLE and ProgramBench; top tier costs about $0.2 per task.
Yue 岳增五
@ZengwuYArchived 10 days agoOriginally posted 2026.10.17
Yue 岳增五 @ZengwuYOriginally posted 2026.10.17
DeepSeek-V4.1-Flash-1017 按官方图表，它在编程基准 DeepSWE v1.1 上拿到 77.0%，比 Opus 5 的 74.0 和 GPT-5.6-Sol 的 73.0 都高；用最高档每个任务大约花 0.2 美元。 不过在 HLE、ProgramBench 这几项上，它还比不过 Opus 5
This post quotes another post: x.com/zengwuy/status/2111376469582856403
x.com/zengwuy/status/2111386348036989126Archived 2026.10.18 16:48