Opus 5.5 is 2.5x cheaper per token, yet costs $6.62 vs $3.89 per task
airesearch12 · x · 2026-09-24
Benchmark Heaven, citing Artificial Analysis data, fact-checked whether Claude Opus 5.5 is really cheaper:
- Per token, Opus 5.5 is 2.5x cheaper than GPT-6 Astra
- Per task it flips: $6.62 vs $3.89 on the AA Intelligence Index
- The culprit is token efficiency: Opus 5.5 emits 119k output tokens per task vs 27k (max effort)
Takeaway: sticker prices mislead — output verbosity can make the "cheaper" model cost more in practice.
Related event: Opus 5.5 Cheaper per Token but Pricier per Task(2 posts)→
More from Models
- User claims MiniMax H3 was trained on The Will Stancil Show after model adds unprompted whistle — Kyrannio · 2026-09-24
- Xiaomi's MiMo V2.6 Pro builds a habit tracker in 62 seconds for under 5 cents — socialwithaayan · 2026-09-24
- Gemini 4 Is Almost Ready, Says New Google DeepMind Chief Koray Kavukcuoglu — The Verge AI · 2026-09-24
- New chat, no refusal: user shows ChatGPT's self-image prompt only blocked in original thread — Todesluke · 2026-09-24
- Open Replications Miss the Secret Sauce: HF Collection Curates Datasets for Jev-Style Models — vanstriendaniel · 2026-09-24
- GPT-6 Astra Fixes a UI That Even Opus 5.5 Couldn't Crack — haltakov · 2026-09-24