PostTrainBench leaderboard
- Claude Fable 5 leads
- 41.8%
- models measured
- 11
- lead over #2
- +5.6 pts
- median
- 31.7%
Leader on PostTrainBench · read October 21
On October 21, 2026, 11 of our catalog's models had a PostTrainBench result, and Anthropic Claude Fable 5 led with 41.8%, ahead of OpenAI GPT-5.6 Sol at 36.2%.
Epoch AI collects its results; its benchmarking hub describes the tasks and links the authors.
Role in the index
PostTrainBench is shown as published and does not feed any of our indexes. Read the methodology
| # | Model | Result | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|
| 1 | Claude Fable 5Anthropic | 41.8% | 79.3 | $50.00 | |
| 2 | GPT-5.6 SolOpenAI | 36.2% | 73.5 | $20.00 | |
| 3 | Claude Opus 5Anthropic | 35.0% | 83.0 | $25.00 | |
| 4 | Claude Opus 4.8Anthropic | 33.8% | 65.0 | $25.00 | |
| 5 | Kimi K3Moonshot AI | 32.0% | 62.3 | $15.00 | |
| 6 | GLM 5.2Z.ai | 31.7% | 61.1 | $4.40 | |
| 7 | Claude Opus 4.7Anthropic | 28.6% | 67.5 | $25.00 | |
| 8 | GPT-5.5OpenAI | 27.2% | 71.3 | $30.00 | |
| 9 | Grok 4.5xAI | 23.4% | 58.6 | $6.00 | |
| 10 | Gemini 3.1 Pro PreviewGoogle | 22.0% | 32.3 | $12.00 | |
| 11 | GPT-5.4OpenAI | 19.0% | 64.4 | $15.00 |
Only models in our catalog are listed, and each appears once with its best published run. 215 rows were read from the source on October 21, 2026, and rows that match no catalog model are left out.