Codex Ultrafast touts 8x faster tokens at 300 tok/s, new ChatGPT Pro plan surfaces
cedric_chee · x · 2026-09-30
- Codex gains an "Ultrafast" mode reportedly delivering up to 8x faster token generation at 300 tok/s, and up to 6x faster in the API.
- A screenshot hints at a new ChatGPT Pro plan bundled with Astra Ultrafast.
- This appears to be an early leak/screenshot share — treat as unconfirmed until OpenAI officially announces.
Related event: GPT-6 Astra Ultrafast Reportedly 8x Faster in Codex(2 posts)→
More from Models
- MiMO ships an agentic critic; early speedups but long-run bias concerns — teortaxesTex · 2026-09-30
- Intern-Decision open models (0.8B–4B) beat Jev on multimodal decision benchmarks — max_paperclips · 2026-09-30
- No reason for OpenAI to keep both $100 and $200 plans, argues Redditor — Over_Sheepherder4503 · 2026-09-30
- Heavy user proposes pairing OAI Pro with a 512GB Mac Studio running GLM-5.3 locally — paperchow · 2026-09-30
- Altman 'Will Think About' More Open Models; Community Points to NVIDIA Nemotron Instead — omarsar0 · 2026-09-30
- "GPT SOL 6.1" Rumor: OpenAI's Weekly Release Cadence Points to a Staggering Base Model — bindureddy · 2026-09-30