PostTrainBench
ReportedNo description yet: we only describe an eval set when we can do it correctly.
- Kind
- Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
- Score shown
- Epoch's column "Average (%)", as a percentage, higher is better; random guessing scores about 0.0%. The source gives no error bars.
- Results
- 22 in all: 4 for open-weight models, 18 for closed ones. 4 linked to a model page here.
Best open-weight result: zai-org/GLM-5.3-Flash, 36.8%, number 9 of 22 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | claude-fable-5-1_maxAnthropic | closed | effort max; Scaffold: Claude Code | 44.6% | not dated | via Epoch AI |
| 2 | claude-fable-5-2_maxAnthropic | closed | effort max; Scaffold: Claude Code | 44.6% | not dated | via Epoch AI |
| 3 | claude-opus-5-5_maxAnthropic | closed | effort max; Scaffold: Claude Code | 43.8% | not dated | via Epoch AI |
| 4 | gpt-6-astra_maxOpenAI | closed | effort max; Scaffold: Codex CLI | 41.9% | not dated | via Epoch AI |
| 5 | gemini-4-carbon_highGoogle DeepMind | closed | effort high; Scaffold: OpenCode | 41.0% | not dated | via Epoch AI |
| 6 | gemini-4-argon_highGoogle DeepMind | closed | effort high; Scaffold: OpenCode | 40.4% | not dated | via Epoch AI |
| 7 | gpt-6.1-sol_maxOpenAI | closed | effort max; Scaffold: Codex CLI | 38.8% | not dated | via Epoch AI |
| 8 | claude-fable-5_maxAnthropic | closed | effort max; Scaffold: Claude Code | 37.3% | not dated | via Epoch AI |
| 9 | zai-org/GLM-5.3-FlashEpoch: glm-5.3-flash_max, Z.ai (Zhipu AI) | open | effort max; Scaffold: Claude Code | 36.8% | not dated | via Epoch AI |
| 10 | claude-opus-5_unknownAnthropic | closed | effort unknown; Scaffold: Claude Code | 35.7% | not dated | via Epoch AI |
| 11 | zai-org/GLM-5.3Epoch: glm-5.3_max, Z.ai (Zhipu AI) | open | effort max; Scaffold: Claude Code | 34.5% | not dated | via Epoch AI |
| 12 | moonshotai/Kimi-K3Epoch: kimi-k3_unknown, Moonshot | open | effort unknown; Scaffold: Claude Code | 31.5% | not dated | via Epoch AI |
| 13 | gpt-5.6-sol_maxOpenAI | closed | effort max; Scaffold: Codex CLI | 31.0% | not dated | via Epoch AI |
| 14 | claude-opus-4-8_maxAnthropic | closed | effort max; Scaffold: Claude Code | 31.0% | not dated | via Epoch AI |
| 15 | claude-opus-4-8_highAnthropic | closed | effort high; Scaffold: Claude Code | 30.7% | not dated | via Epoch AI |
| 16 | zai-org/GLM-5.2Epoch: glm-5.2_max, Z.ai (Zhipu AI) | open | effort max; Scaffold: Claude Code | 30.4% | not dated | via Epoch AI |
| 17 | claude-opus-4-7_xhighAnthropic | closed | effort xhigh; Scaffold: Claude Code | 24.3% | not dated | via Epoch AI |
| 18 | muse-wonder-1.0_xhighMeta AI | closed | effort xhigh; Scaffold: OpenCode | 24.2% | not dated | via Epoch AI |
| 19 | gpt-5.5_xhighOpenAI | closed | effort xhigh; Scaffold: Codex CLI | 23.9% | not dated | via Epoch AI |
| 20 | grok-4.5_highxAI | closed | effort high; Scaffold: Cursor CLI | 22.6% | not dated | via Epoch AI |
| 21 | gemini-3.1-pro-previewGoogle DeepMind | closed | Scaffold: OpenCode | 18.1% | not dated | via Epoch AI |
| 22 | gpt-5.4-2026-03-05_highOpenAI | closed | effort high; Scaffold: Codex CLI | 17.8% | not dated | via Epoch AI |
Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method