Baseten
GLM 5.2
Made by Z.ai (Zhipu AI). Released 2026-06-16. Accessibility not recorded. Cheapest standard rate we hold: $4.40 per million output tokens, no context split.
What it costs
| Tier | Kind | In | Out | Cached in | From |
|---|---|---|---|---|---|
| standard | text | $1.40 | $4.40 | $0.30 | *registry-single |
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 | 0.1±0.02 | 2026-08-10 | none |
| chess puzzles | Epoch AI | 0.1±0.03 | 2026-08-10 | low |
| chess puzzles | Epoch AI | 0.2±0.04 | 2026-06-17 | max |
| ebr bench | Epoch AI | 0.1 | 2026-06-29 | max |
| frontiermath tier 4 v2 | Epoch AI | 0.3±0.07 | 2026-06-19 | max |
| frontiermath tiers 1 3 v2 | Epoch AI | 0.4±0.03 | 2026-08-29 | none |
| frontiermath tiers 1 3 v2 | Epoch AI | 0.5±0.03 | 2026-08-29 | low |
| frontiermath tiers 1 3 v2 | Epoch AI | 0.6±0.03 | 2026-06-19 | max |
| gpqa diamond | Epoch AI | 0.9±0.02 | 2026-08-10 | low |
| gpqa diamond | Epoch AI | 0.7±0.03 | 2026-08-10 | none |
| gpqa diamond | Epoch AI | 0.9±0.02 | 2026-06-24 | max |
| mystery game puzzles | Epoch AI | 0.1±0.04 | 2026-08-27 | medium |
| mystery game puzzles | Epoch AI | 0.2±0.04 | 2026-08-27 | minimal |
| mystery game puzzles | Epoch AI | 0.2±0.04 | 2026-08-27 | low |
| mystery game puzzles | Epoch AI | 0.2±0.04 | 2026-08-27 | none |
| otis mock aime 2024 2025near its ceiling | Epoch AI | 0.8±0.06 | 2026-08-10 | low |
| otis mock aime 2024 2025near its ceiling | Epoch AI | 0.3±0.07 | 2026-08-10 | none |
| otis mock aime 2024 2025near its ceiling | Epoch AI | 0.9±0.04 | 2026-06-25 | max |
| simpleqa verified | Epoch AI | 0.3±0.02 | 2026-08-27 | max |
| swe bench verified | Epoch AI | 0.8±0.02 | 2026-06-25 | max |
| ale bench | collected by Epoch AI | 1047 | — | high |
| ale bench | collected by Epoch AI | 1010 | — | max |
| apex agents | collected by Epoch AI | 0.5 | — | unknown |
| arc aginear its ceiling | collected by Epoch AI | 0.8 | — | unknown |
| arc agi 2 | collected by Epoch AI | 0.2 | — | unknown |
| critpt | collected by Epoch AI | 0.2 | — | max |
| critpt | collected by Epoch AI | 0.0 | — | none |
| cursorbench | collected by Epoch AI | 0.5 | — | high |
| cursorbench | collected by Epoch AI | 0.6 | — | max |
| deepswe | collected by Epoch AI | 0.4 | — | mini-swe-agent |
| dtbenchnear its ceiling | collected by Epoch AI | 0.9 | — | max |
| frontiercode | collected by Epoch AI | 0.2 | — | mini-swe-agent |
| gbaeval | collected by Epoch AI | 0.0 | — | unknown |
| lmca | collected by Epoch AI | 45.8 | — | max |
| posttrainbench | collected by Epoch AI | 0.3 | — | Claude Code |
| proofbench | collected by Epoch AI | 0.3 | — | max |
| scicode | collected by Epoch AI | 0.4 | — | none |
| scicode | collected by Epoch AI | 0.5 | — | max |
| simplebench | collected by Epoch AI | 0.6 | — | unknown |
| surface evolver bench | collected by Epoch AI | 0.6 | — | high |
| vending bench 2 | collected by Epoch AI | 8314 | — | unknown |
| webdev arena | collected by Epoch AI | 1593 | — | max |
| weirdml | collected by Epoch AI | 0.7 | — | high |
| weirdml | collected by Epoch AI | 0.7 | — | max |
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 Baseten