xAI
grok-4.5
Made by xAI. Released 2026-07-08. Accessibility not recorded. Cheapest standard rate we hold: $6 per million output tokens, at < 200k prompt tokens context.
What it costs
| Tier | Kind | In | Out | Cached in | From |
|---|---|---|---|---|---|
| standard | text · < 200k prompt tokens | $2 | $6 | $0.30 | seller |
| standard | text · ≥ 200k prompt tokens | $4 | $12 | $0.60 | 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 | 0.4±0.05 | 2026-07-08 | high |
| frontiermath tier 4 v2 | Epoch AI | 0.2±0.07 | 2026-07-09 | high |
| frontiermath tiers 1 3 v2 | Epoch AI | 0.6±0.03 | 2026-07-09 | high |
| gpqa diamond | Epoch AI | 0.9±0.01 | 2026-07-08 | high |
| otis mock aime 2024 2025near its ceiling | Epoch AI | 1.0±0.01 | 2026-07-08 | high |
| simpleqa verified | Epoch AI | 0.5±0.02 | 2026-08-27 | high |
| ale bench | collected by Epoch AI | 1309 | — | high |
| apex agents | collected by Epoch AI | 0.5 | — | unknown |
| arc aginear its ceiling | collected by Epoch AI | 0.9 | — | medium |
| arc aginear its ceiling | collected by Epoch AI | 0.9 | — | high |
| arc aginear its ceiling | collected by Epoch AI | 0.8 | — | low |
| arc agi 2 | collected by Epoch AI | 0.5 | — | medium |
| arc agi 2 | collected by Epoch AI | 0.5 | — | high |
| arc agi 2 | collected by Epoch AI | 0.3 | — | low |
| blueprint bench 2 | collected by Epoch AI | 0.3 | — | unknown |
| critpt | collected by Epoch AI | 0.2 | — | high |
| deepswe | collected by Epoch AI | 0.5 | — | mini-swe-agent |
| dtbenchnear its ceiling | collected by Epoch AI | 1.0 | — | high |
| frontiercode | collected by Epoch AI | 0.4 | — | grok-build |
| gbaeval | collected by Epoch AI | 0.7 | — | unknown |
| gdp pdf | collected by Epoch AI | 0.1 | — | high |
| lmca | collected by Epoch AI | 45.2 | — | high |
| posttrainbench | collected by Epoch AI | 0.2 | — | Cursor CLI |
| proofbench | collected by Epoch AI | 0.3 | — | high |
| scicode | collected by Epoch AI | 0.5 | — | high |
| simplebench | collected by Epoch AI | 0.7 | — | unknown |
| surface evolver bench | collected by Epoch AI | 0.7 | — | high |
| vending bench 2 | collected by Epoch AI | 3887 | — | unknown |
| webdev arena | collected by Epoch AI | 1556 | — | unknown |
| weirdml | collected by Epoch AI | 0.5 | — | unknown |
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 xAI