Gemini 2.5 Pro
Made by Google DeepMind. Released 2025-06-17. Accessibility not recorded. Cheapest standard rate we hold: $10 per million output tokens, no context split.
What it costs
| Tier | Kind | In | Out | Cached in | From |
|---|---|---|---|---|---|
| batch | text | $0.63 | $5 | — | seller |
| flex | text | $0.63 | $5 | — | seller |
| priority | text | $2.25 | $18 | — | seller |
| standard | text | $1.25 | $10 | — | seller |
per million tokens, US dollars. * marks a rate from a single registry or two that disagree — the best available number rather than a confirmed one. Every other row is confirmed against a seller or a licensed index.
What it scores, by house
| Benchmark | House | Score | Evaluated | Scaffold |
|---|---|---|---|---|
| chess puzzles | Epoch AI | 0.2±0.04 | 2025-12-08 | — |
| frontiermath | Epoch AI | 0.1±0.02 | 2025-11-24 | — |
| frontiermath tier 4 | Epoch AI | 0.0±0.03 | 2025-07-03 | — |
| frontiermath tier 4 v2 | Epoch AI | 0.0±0.00 | 2026-06-11 | — |
| frontiermath tiers 1 3 v2 | Epoch AI | 0.2±0.03 | 2026-06-11 | — |
| gpqa diamond | Epoch AI | 0.9±0.02 | 2025-11-16 | — |
| otis mock aime 2024 2025near its ceiling | Epoch AI | 0.8±0.05 | 2025-11-16 | — |
| swe bench verified | Epoch AI | 0.6±0.02 | 2026-02-13 | — |
| ale bench | collected by Epoch AI | 786 | — | 32K |
| algotune | collected by Epoch AI | 1.5 | — | — |
| apex agents | collected by Epoch AI | 0.2 | — | — |
| arc aginear its ceiling | collected by Epoch AI | 0.3 | — | 8K |
| arc aginear its ceiling | collected by Epoch AI | 0.4 | — | 16K |
| arc aginear its ceiling | collected by Epoch AI | 0.4 | — | 32K |
| arc agi 2 | collected by Epoch AI | 0.0 | — | 16K |
| arc agi 2 | collected by Epoch AI | 0.0 | — | 32K |
| arc agi 2 | collected by Epoch AI | 0.0 | — | 8K |
| critpt | collected by Epoch AI | 0.0 | — | — |
| deepresearchbench | collected by Epoch AI | 0.4 | — | — |
| dtbenchnear its ceiling | collected by Epoch AI | 0.8 | — | — |
| forecastbench | collected by Epoch AI | 60.4 | — | — |
| gdpval | collected by Epoch AI | 0.2 | — | — |
| gso | collected by Epoch AI | 0.0 | — | OpenHands |
| lech mazur writing | collected by Epoch AI | 8.4 | — | — |
| lmca | collected by Epoch AI | 34.8 | — | — |
| scicode | collected by Epoch AI | 0.4 | — | — |
| spatialviz bench | collected by Epoch AI | 0.4 | — | — |
| terminalbench | collected by Epoch AI | 0.3 | — | Mini-SWE-Agent |
| terminalbench | collected by Epoch AI | 0.3 | — | Terminus 2 |
| terminalbench | collected by Epoch AI | 0.2 | — | Gemini CLI |
| terminalbench | collected by Epoch AI | 0.2 | — | OpenHands |
| vending bench 2 | collected by Epoch AI | 574 | — | — |
| webdev arena | collected by Epoch AI | 1226 | — | — |
| weirdml | collected by Epoch AI | 0.5 | — | 16K |
These scores are not comparable between houses and are not added up. A score is dated by when it was evaluated, not when the model was released, and a score without its scaffold is not a measurement — which is why both are printed.
Every rate we hold · Why there is no single best model · Everything from Google