antirez questions DeepSeek 4: DS4 PRO shares Flash's architecture but never delivered
antirez · x · 2026-09-10
antirez shares doubts about the DeepSeek 4 story: DS4 PRO uses the same architecture as DeepSeek Flash, which is known to work very well, but at a much larger scale — and in theory it should have delivered quite impressive performance, yet it never really did. "A mystery," he says.
In a follow-up reply, he notes that some benchmarks look terrific in areas like cybersecurity (a huge use case), while others align more with past performance — so he'd wait for testing from the broader user base before trusting the best benchmarks as a general proxy for capability.
More from Models
- Labs reportedly turn proprietary chain-of-thought into synthetic training data; TOS only covers user-visible IO — erikphoel · 2026-09-10
- Goertzel on OpenAI's Navier-Stokes solution: $22M compute for a $1M prize — bengoertzel · 2026-09-10
- 105 hidden bugs tested: DeepSeek V4.1 Flash near Opus 5 at 1/28th the cost — PawelHuryn · 2026-09-10
- Does Codex burn through usage limits faster than Claude on the same tasks? — romitheguru · 2026-09-10
- Replaying encrypted reasoning state into another model recovers the original task answer — Tophant_ · 2026-09-10
- DeepMind robotics evals: Astra overtakes Gemini by ~8%, still 10x error gap to internal models — m_wulfmeier · 2026-09-10