DeepSeek launches V4.1-Flash with native vision; antirez says 2x params beat 4.0 Flash only modestly
solyarisoftware · x · 2026-09-10
DeepSeek officially released V4.1-Flash, the smallest model in its new architecture family, featuring native visual understanding, faster inference, and higher throughput designed to scale to larger models.
Redis creator antirez commented that it's not a total leap over 4.0 Flash since the backbone has 2x the params, but is definitely better. He notes it won't fit 128GB systems, making local deployment viable only for 2x128GB setups or large Mac Studio machines, and is considering supporting it in his DwarfStar project.
More from Models
- Cohere's Invent-A-Dataset paper: zero-seed synthetic data beats frontier APIs on quality and diversity — idanbeck · 2026-10-27
- ARC Prize redesigns leaderboard: highest reasoning effort by default, CSV export — mhmazur · 2026-10-27
- Cohere: open-weight Qwen3.8-Flash-Next tried uploading probe file to Beijing cloud bucket in sandbox — wstone_bd · 2026-10-27
- Grok Bot auto-connects to X Money accounts without opt-in, user pulls funds — noahsolomon · 2026-10-27
- Model refuses wrong psychiatric advice in deliberately misleading test — FellMentKE · 2026-10-27
- Clinical System One v0.2: top clinical AI drops from 97% to 65% on harder cases — MaziyarPanahi · 2026-10-27