Token pricing is meaningless now: Astra cheaper per task than Gemini 3.8 Flash
stevenheidel · x · 2026-09-04
OpenAI's Steven Heidel says token pricing is effectively meaningless: GPT-6 Astra sits alone on the pareto frontier of cost efficiency due to extreme token efficiency, costing less per task than Gemini 3.8 Flash — a model 13x cheaper per token. His advice: measure your costs per task, not per token.
Related event: GPT-6 Astra Cheaper Per Task Despite Higher Token Price(2 posts)→
More from Models
- Google confirms Gemini 3.8 Flash in AI Mode drops citations and links, fix on the way — gaganghotra_ · 2026-09-04
- Quick Question: Does GPT-6 Include HuggingFace Access? — gordic_aleksa · 2026-09-04
- MiniMax Video Upscaler Tip: Skip the 60-120s 'Quality' Prompt Expansion Mode — altryne · 2026-09-04
- "Imagine believing in benchmarks in 2026": AI circle mocks leaderboard worship — vasuman · 2026-09-04
- The real interactivity test: learning a new language purely by talking to an LLM — akbirthko · 2026-09-04
- 753B model 'thinks', 4B model writes: latent-space handoff claimed to be 20x faster — burny_tech · 2026-09-04