All models get ultrafast: ~300 tps, 8x faster but 6x more expensive
cneuralnetwork · x · 2026-09-30
OpenAI's Ultrafast tier is now available across all models, delivering roughly 8x faster inference at 300 tokens per second — at about 6x the price. Aimed at latency-critical workloads, it forces users to weigh speed against cost.
More from Models
- European users complain Dots feature missing despite paying the same price — AweVR · 2026-09-30
- OpenAI rolls out Codex cloud environments to Plus, Pro, Business and Enterprise — OpenAIDevs · 2026-09-30
- Theo on Sol 6.1: great model and incredible value, but is it as good as Opus 5.5? — heyneighbor · 2026-09-30
- OpenAI launches Dots, always-on agents powered by GPT-6 Astra that work 24/7 across 4,000+ apps — OpenAI · 2026-09-30
- Heavy user's math: mixing $100 Claude, $200 Codex, $30 Grok beats a single max plan — codtv132 · 2026-09-30
- Price of intelligence now halves every 2 weeks, leaving enterprise budgets guessing — gabriel1 · 2026-09-30