You pay for invisible thinking tokens: why closed LLM pricing can't be verified
burkov · x · 2026-09-19
- Andriy Burkov points out a hidden billing fact: closed-weight LLM providers charge for thinking tokens but never send them to the caller.
- So listed prices (e.g., $2/1M input, $10/1M output) could be artificially underpriced, with providers recovering costs by billing "thinking tokens" they may never have actually generated.
- Users have no way to verify the number of thinking tokens they're charged for—pricing transparency rests entirely on the provider's honesty.
Related event: Burkov questions opaque thinking-token billing in closed LLMs(2 posts)→
More from Models
- Users slam OpenAI's Advanced Voice Mode: non-English support feels unfinished — Angaisb_ · 2026-09-19
- Yandex releases AliceAI-80B-A3B base model trained from scratch with 262K context under Apache 2.0 — cephaloform · 2026-09-19
- Muse Spark 1.3 Gets Cheaper Contributor-Tier Optimization With Only 1-2% Benchmark Variance — alexandr_wang · 2026-09-19
- Unreleased Tencent Hunyuan 3.5 Spotted in Early Access on OnSolo, Image Quality Impresses — HeyAmit_ · 2026-09-19
- Auto-Formalizing a 76-Page Paper With Opus 5 High Would Take ~40 Days — kfountou · 2026-09-19
- Open-weight models now take 56% of production token volume, per Vercel index — cramforce · 2026-09-19