V4.1 Flash is ~3x faster than V4-Flash and smarter per token, shows sensible self-debugging
teortaxesTex · x · 2026-09-08
Blogger teortaxesTex shares hands-on observations of a new model version: V4.1 Flash is roughly 3x faster than V4-Flash while seeming smarter per token. It likes to take lots of images and reason hard about them, and shows fairly reasonable ideas about debugging its own work — with 'Visual Primitives' repeatedly delivering.
Related event: V4.1 Flash benchmarks about 3x faster than V4-Flash(2 posts)→
More from Models
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27
- ChatGPT's crawler sample code strips out all Schema markup — VeryWellVersed · 2026-10-27