Why are some GPT-5 API requests costing 3x more? Hidden reasoning tokens suspected
Swimming-Skirt3870 · reddit · 2026-10-24
A developer processing documents with the GPT-5 API noticed identical-length documents costing 2-3x more per request. Input size doesn't explain it — some requests burn far more output tokens, likely hidden reasoning tokens rather than the actual response. This makes batch cost estimation unreliable. They ask how production users separately track reasoning vs. response tokens to audit per-request cost.
More from Models
- Abacus AI to launch SuperComputer: text prompt to custom-trained LLM in an hour — bindureddy · 2026-10-24
- One-prompt Minecraft build of an RTX 5090 sparks model-vs-Opus 4.5 poll — adonis_singh · 2026-10-24
- Abliterated DeepSeek running on a Pi is the AI power-user flex of the day — EXM7777 · 2026-10-24
- Tesla engineer: Cybercab camera tuning done with Grok, coding fixed-function hardware directly — yunta_tsai · 2026-10-24
- VulnPR-100: DeepSeek V4.1 Flash finds 54 vulns for $7.53, crushing rivals on cost — teortaxesTex · 2026-10-24
- Kimi K3.1 Pro costs $1.03 median per task but lands just off Agent Arena's Pareto frontier — arena · 2026-10-24