PostTrainBench
Practical engineering evaluation of ML post-training. This covers results whose version is not stated in the source (v1.2 is kept separate as posttrainbench-1-2).
- Category
- Other
- Unit
- Percent (%)
- Higher is better
- Comparison groups
- 3total
- All model rankings (records) · Ranked results: 3
All model rankings (records)
These rankings are recalculated by comparison group including legacy models and models unavailable in Claude Code / Codex. They differ from the rankings above because the set of models differs.
Show rankings including legacy and unavailable models (entries: 3)
official results (Gemini 4 Argon model card)
| Rank | Model | effort | Score | Source cost (metric) | measured/reported by | source |
|---|---|---|---|---|---|---|
| 1 | Gemini 4 Argon | not stated | 45.3% | — | Gemini 4 Argon model cardRetrieved 2026-11-12 · Vendor-reported value. Self-reported by Google (model card, as of 2026-10). thinking setting not statedSource excerptPostTrainBench | 45.3% |
official results (Kimi K3 (Hugging Face model card, moonshotai))
| Rank | Model | effort | Score | Source cost (metric) | measured/reported by | source |
|---|---|---|---|---|---|---|
| 1 | Kimi K3 | Max | 36.6% | — | Moonshot AI | Kimi K3 (Hugging Face model card, moonshotai)Retrieved 2026-11-13 · Vendor-reported value (self-reported by Moonshot AI). Table note: All Kimi K3 results are obtained with reasoning effort set to 'max' and temperature = 1.0. Official Harbor implementation; maximum effort; average of 3 runs on H20 GPU instead of H100; Claude Code harness.Source excerptPostTrainBench | 36.6Location HF model card "3. Evaluation Results" table, PostTrainBench row, Kimi K3 (max) column (leftmost numeric column). Supporting footnote: Coding / PostTrainBench |
official results (GLM-5.3 (Hugging Face model card, zai-org))
| Rank | Model | effort | Score | Source cost (metric) | measured/reported by | source |
|---|---|---|---|---|---|---|
| 1 | GLM-5.3 | Max | 39.8% | — | Z.ai | GLM-5.3 (Hugging Face model card, zai-org)Retrieved 2026-11-13 · Vendor-reported value (self-reported by Z.ai). Card Note: the default for reasoning_effort is max; use the default max to reproduce the benchmarks. Claude Code 2.1.207; max effort level; weighted average of 3 runs. The third-party API usage check was replaced with an LLM agent.Source excerptPostTrainBench | 39.8Location HF model card "Benchmark" table, PostTrainBench row, GLM-5.3 column (leftmost numeric column). Basis: Footnotes, PostTrainBench |