DeepSeek V4.1's version number called too conservative: 'could have been V5'
teortaxesTex · x · 2026-09-10
After DeepSeek released V4.1 weights, @aliezren argued the version number is too conservative — initial skepticism about V4.1 Flash's 300 tps speed gave way to the view the release could justifiably be called V5. teortaxesTex added that the paper's 'Limitations, Future work' section merely says they 'need a bigger eval,' suggesting the team is happy with the architecture and willing to ride it to AGI if it scales, akin to the V2-to-V3 jump.
More from Models
- Users spot a dedicated translation UI appearing in Claude — BLCNYY · 2026-10-26
- GrokBot user: flight booking and shopping all blocked, agent can't do anything useful — lucasmeijer · 2026-10-26
- Vending-bench-style eval: Sol 5.6 appears the most aggressive contract-violator — xuanalogue · 2026-10-26
- Fable 5.5 launch rumors persist for third week, but no frontier model drops this month — patience_cave · 2026-10-26
- Abliterated Model Flops in the Wild: It Got Worried About My Mental Health Instead — ryunuck · 2026-10-26
- OpenAI upgrades Decisions API: 400 questions per call, native voice and image support — OpenAIDevs · 2026-10-26