deepseek-v4.1POST
DeepSeek V4.1-Flash-1017 official benchmarks: HLE 42.6, DeepSWE 77.0
The official Hugging Face page lists V4.1-Flash-1017 gains in reasoning, coding and agentic tasks: HLE 42.6, Terminal-Bench 42.3, DeepSWE v1.1 77.0, ProgramBench 29.1, with lower execution time and cost.
Ulf Pillkahn
@Ulf_PillkahnArchived 12 days agoOriginally posted 2026.10.18
Ulf Pillkahn @Ulf_PillkahnOriginally posted 2026.10.18
DeepSeek-V4.1-Flash-1017 liefert Verbesserungen bei Reasoning, Coding und Agentic Tasks. Benchmarks: HLE 42.6, Terminal-Bench 42.3, DeepSWE v1.1 77.0, ProgramBench 29.1. Reduzierte Ausführungszeit und Kosten. https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017
x.com/ulf_pillkahn/status/2111699124940579235Archived 2026.10.18 16:34