deepseek-v4.1POST
DeepSeek V4.1 Flash 1017 benchmarks: DeepSWE 77.0 near Argon, HLE 7–15 behind
The author reads the vendor table: DeepSWE 77.0 is close to Gemini 4 Argon's 77.9 and GPT-6.1 Sol's 75.2, while HLE 42.6 ties GLM-5.3 and Qwen3.8 Max but trails four overseas models by 7 to 15 points.
Kiwi Wu🏂🐱🐦⬛
@maisui99Archived 9 days agoOriginally posted 2026.10.19
Kiwi Wu🏂🐱🐦⬛ @maisui99Originally posted 2026.10.19
对比了一下数据 DeepSeek-V4.1-Flash-1017 在厂商表里最高的几项（DeepSWE 77.0、AutomationBench 60.5）对手是上一代海外模型（GPT-5.6 Sol、Opus 5）。能和海外最新模型对齐的只有 DeepSWE 和 HLE 两行：DeepSWE 77.0 与 Argon 77.9、GPT-6.1 Sol 75.2 接近；HLE 42.6 与 GLM-5.3、Qwen3.8 Max 持平，比海外四款低 7 到 15。
This post quotes another post: x.com/deepseek_ai/status/2111375640717033777
x.com/maisui99/status/2112020526231355546Archived 2026.10.19 12:02