Benchmark
PostTrainBench
PostTrainBench evaluates a model's ability to autonomously post-train base models. Given pretrain-only base models, the agent must complete the full…
- Modality
- text
- Categories
- reasoning, agents, code, systems
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 6
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| GLM-5.2 | Z.ai | 0.343 | 2026-06-16 | evidence |
| GLM-5.3 | Z.ai | 0.398 | 2026-08-14 | evidence |
| Kimi K3 | Moonshot AI | 0.366 | 2026-07-16 | evidence |
| MiniMax M3 | MiniMax | 0.371 | 2026-06-01 | evidence |
| Seed 2.1 Pro | ByteDance | 0.165 | 2026-06-24 | evidence |
| Seed 2.1 Turbo | ByteDance | 0.183 | 2026-06-24 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.