OpenAI's subsidy shift rewrites the score-vs-cost Pareto frontier, analysis finds
RexDouglass · x · 2026-10-30
Adam Holter's mypriors.com analysis ("Score vs cost per task", with resets counted) argues OpenAI's new subsidization changes the value Pareto frontier: unsubsidized, its flagship is roughly in line with Claude, but the subsidy pulls its per-task cost curve well below competitors.
On the AA-II composite (10 benchmarks incl. AA-Briefcase v1.1, FrontierCode 2.0-AA and Terminal-Bench-Science-AA at 15% each), scores are Gemini 4 Carbon 59.4, Opus 5.5 58.2, GPT-6 Astra 56.8. Key takeaway: nominal prices and real subscription economics now diverge, so model comparisons must count reset/limit costs to see the true frontier.
More from Models
- Kimi K3.1 Pro leak teases open weights by Nov 4, 2026 as progress index reportedly tops Fable 5 — teortaxesTex · 2026-10-30
- Users complain Opus 5.5 writes 'illegible rushed technical speak' and won't fix it — msalbergo · 2026-10-30
- DeepSeek V4.1 Flash tops open-model usage among research agents, index data shows — AkariAsai · 2026-10-30
- Dev claims Gemini Argon makes 'researcher-level jumps' other models miss as 'carbon' looms — vitaliychiley · 2026-10-30
- Meta's Alexandr Wang touts Muse: 6.7x more concise than ChatGPT at answering the same question — alexandr_wang · 2026-10-30
- Phonic tops VAmoS benchmark for agentic speech-to-speech models — graceisford · 2026-10-30