tokens per RL trajectory for reasoning/agentic tasks — the rollout that dominates cost
10K–100K+observed
| Value kind | observed — Reported measurements, counts, and specifications keep the precision and scope stated by their source; an exact specification is not treated as a range. |
|---|---|
| As of | 2026 |
| Source | reasoning-model RL reports · Abstract; reasoning trajectories are 500–30K+ tokens and agentic trajectories 100K–1M tokens |
| Review | checking…review by 2026-08-25 · fast cadence |
| Recorded changes | last 2026-07-10 · 2 revisions tracked |
| Claim id | tokens-per-rl-trajectory-for-reasoning-agentic |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.