User claims DeepSeek kept V4 Pro online under pressure, slowing next-gen training
teortaxesTex · x · 2026-09-13
X user teortaxesTex complains (unverified) that Chinese users pressured DeepSeek into not retiring V4 Pro, calling it a "gigantic HBM hog" that burns thousands of rollouts/sec — compute that would otherwise go to training the next 4.2 model — and may force peak/off-peak scheduling or another price hike. The post highlights the real tension between serving flagship models and funding frontier training, and how user pressure can shape model retirement decisions.
More from Infra
- Dev picks LFM2.5-2.6B and MiniCPM5-2B as favorite edge-device models — reach_vb · 2026-09-13
- Prompt caching can inflate your LLM bill: check write premiums, TTL and actual reuse — gethackteam · 2026-09-13
- Speculative decoding: the trick behind Google's 2-3x faster LLM inference in production — hongyangzh · 2026-09-13
- TRL v1.13 ships long-context training: 1M+ token sequences on a single 8-GPU H100 node — SergioPaniego · 2026-09-13
- SWE builds a 3x RTX 3090, 64GB VRAM local coding workstation: full parts and pitfalls — trytoinfect74 · 2026-09-13
- Enterprise AI costs 10-20x traditional systems; private cloud may win big — DavidLinthicum · 2026-09-13