Benchmarks · reasoning
PostTrainBench v1
Uses the archived v1 weighted aggregate; subtasks do not count as separate benchmarks. Score: weighted aggregate score. Collected by Jake Boggs' benchmark dashboard.
13 models · 14 results · data as of 2026-10-29 · posttrainbench.com ↗
Leaderboard
GLM 5.2 max
All results
14 rows| # | Model | Effort | PostTrainBench v1 |
|---|---|---|---|
| 1 | GLM 5.2 | max | 34.3% |
| 2 | Opus 4.8 | max | 34.1% |
| 3 | Opus 4.7 | max | 28.6% |
| 4 | GPT-5.5 | xhigh | 25.0% |
| 5 | Opus 4.6 | max | 24.8% |
| 6 | Gemini 3.1 Pro | high | 21.6% |
| 7 | GPT-5.2 | xhigh | 21.4% |
| 8 | GPT-5.4 | high | 20.2% |
| 9 | GPT-5.1 Codex | max | 19.7% |
| 10 | Gemini 3 Pro | high | 18.1% |
| 11 | GPT-5.3 Codex | high | 17.8% |
| 12 | GPT-5.2 Codex | xhigh | 17.2% |
| 13 | Opus 4.5 | high | 17.1% |
| 14 | GPT-5.3 Codex | medium | 13.8% |