Long Reasoning Task Consumes 22k Tokens for 3k Output
generativist · x · 2026-08-15
Simon Willison shared test data from an extremely long reasoning task.
Key Stats:
- Duration: Nearly 21 minutes.
- Token Usage: Used 22,276 reasoning tokens to produce only 3,223 output tokens.
This data highlights the massive compute cost of high-end models on complex reasoning tasks, where the "thinking" cost far outweighs the final output generation cost.
More from Infra
- CoreWeave's debt costs falling: DDTL spreads tightened from S+9.62% to S+2.25% — BenBajarin · 2026-10-16
- SoftBank lines up US IPOs for SB Energy and AI robotics spinoff Roze, valuations top $50B — nordicinst · 2026-10-16
- TRL reproduces Modal's mixed-precision async RL: plain GRPO collapses, SC+MIS holds 0.72 — QGallouedec · 2026-10-16
- Hugging Face scales Indic pretraining with 2.8T tokens and an open data pipeline — Hugging Face Blog · 2026-10-16
- NVIDIA Vera Rubin NVL72 hits up to 5.7x GB300 throughput per MW on agent benchmark — ArtificialAnlys · 2026-10-16
- 14 open-source LLM observability tools to know in 2026, from Langfuse to Laminar — TheTuringPost · 2026-10-16