DeepSeek V4.1-Flash leaker accused of being a bot: 'not a single MLA layer'
teortaxesTex · x · 2026-10-18
AI circle account teortaxesTex accuses a prolific DeepSeek leaker of being an automated fake: 'There is not a single MLA layer in V4.1-Flash,' calling for punishment of automated verified accounts.
- The quoted post claimed that even if DeepSeek V4.1-Flash scores 43-45, the real divergence is Intelligence-per-Dollar & Tokens-per-Watt: Meta's Muse Spark Open reportedly needs an 816B footprint (42B active) to scratch 39, while DeepSeek's moat is algorithmic compression via MLA and fine-grained routing
- The leak claims are unverified; the post itself is a callout of the leak chain
More from Models
- Perplexity's decision model wins Catan tournament at under half a cent per game — denisyarats · 2026-10-18
- DeepSeek V4.1 Flash 1017 aces 10-task coding benchmark at under 5 cents per task — Fabulous_Pollution10 · 2026-10-18
- Scaling01 Slams DeepSeek's New Model as 'Can't Ship Good Models Anymore', Drama Ensues — teortaxesTex · 2026-10-18
- From 4K to 1M tokens: context windows grew 250x in four years since GPT-3.5 — khademinori · 2026-10-18
- Ollama rolls out updated model to DeepSeek-V4.1-Flash users — ollama · 2026-10-18
- DeepSeek models reason about the grader on LisanBench, suggesting RL on the benchmark — scaling01 · 2026-10-18