'Give me 1000 tok/s models': devs say speed, not intelligence, is the bottleneck
DanielLockyer · x · 2026-10-19
Sam Whitmore argues what he wants most from next-gen models is speed: models are 'plenty good now,' and he wants true interactivity — saying 'let's try that split vertical instead of horizontal' and seeing it under a second later. Daniel Lockyer echoes with a request for 1000 tok/s models. The shared take: latency, not intelligence, is becoming the main bottleneck in AI product experience.
More from Models
- ChatGPT web can actually listen to audio, but the same model can't via API or Codex — Possible_Statement84 · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- TLA+ expert corrects Grok: TLC only exhausts finite state spaces, not all behaviors — tianyin_xu · 2026-10-23
- Open LLM cost calculator shows Arabic tokenization and 6x price gaps across models — Moham-saed · 2026-10-23
- OpenAI blocks manual reset with 13% usage left: 'no reset needed right now' — AccomplishedSugar490 · 2026-10-23
- Meta's 816B Muse Spark Open scores 39, losing to 180B Qwen3.8-Flash-Next at 40 — BTA_Labs · 2026-10-23