FrontierSWE
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.
- Link
- https://www.frontierswe.com/
- Score shown
- Epoch's column "Score", as a percentage, higher is better; random guessing scores about 0.0%. The source gives no error bars.
- Results
- 25 in all: 5 for open-weight models, 20 for closed ones. 5 linked to a model page here.
Best open-weight result: zai-org/GLM-5.3, 30.2%, number 13 of 25 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | claude-fable-5-2_maxAnthropic | closed | effort max; Harness: proximus | 69.9% | not dated | frontierswe.com |
| 2 | gpt-6-astra_maxOpenAI | closed | effort max; Harness: proximus | 65.5% | not dated | frontierswe.com |
| 3 | claude-opus-5-5_maxAnthropic | closed | effort max; Harness: proximus | 62.3% | not dated | frontierswe.com |
| 4 | claude-sonnet-5-5_maxAnthropic | closed | effort max; Harness: proximus | 61.9% | not dated | frontierswe.com |
| 5 | gemini-4-carbon_highGoogle DeepMind | closed | effort high; Harness: proximus | 61.2% | not dated | frontierswe.com |
| 6 | claude-fable-5-1_maxAnthropic | closed | effort max; Harness: proximus | 56.3% | not dated | frontierswe.com |
| 7 | gemini-4-argon_highGoogle DeepMind | closed | effort high; Harness: proximus | 55.0% | not dated | frontierswe.com |
| 8 | gemini-4-hydrogen_highGoogle DeepMind | closed | effort high; Harness: proximus | 53.1% | not dated | frontierswe.com |
| 9 | claude-opus-5_maxAnthropic | closed | effort max; Harness: proximus | 52.0% | not dated | frontierswe.com |
| 10 | claude-fable-5_maxAnthropic | closed | effort max; Harness: proximus | 47.0% | not dated | frontierswe.com |
| 11 | claude-haiku-5-5_maxAnthropic | closed | effort max; Harness: proximus | 43.8% | not dated | frontierswe.com |
| 12 | gpt-5.6-sol_maxOpenAI | closed | effort max; Harness: proximus | 32.2% | not dated | frontierswe.com |
| 13 | zai-org/GLM-5.3Epoch: glm-5.3_max, Z.ai (Zhipu AI) | open | effort max; Harness: proximus | 30.2% | not dated | frontierswe.com |
| 14 | grok-4.7_xhighxAI | closed | effort xhigh; Harness: proximus | 29.5% | not dated | frontierswe.com |
| 15 | moonshotai/Kimi-K3Epoch: kimi-k3_max, Moonshot | open | effort max; Harness: proximus | 25.9% | not dated | frontierswe.com |
| 16 | muse-spark-1.3_maxMeta AI | closed | effort max; Harness: proximus | 25.8% | not dated | frontierswe.com |
| 17 | grok-4.6_xhighxAI | closed | effort xhigh; Harness: proximus | 25.3% | not dated | frontierswe.com |
| 18 | thinkingmachines/Inkling-2-SmallEpoch: inkling-2-small_xhigh, Thinking Machines | open | effort xhigh; Harness: proximus | 25.1% | not dated | frontierswe.com |
| 19 | gemini-3.7-flash_highGoogle DeepMind | closed | effort high; Harness: proximus | 20.3% | not dated | frontierswe.com |
| 20 | gemini-3.8-flash_highGoogle DeepMind | closed | effort high; Harness: proximus | 19.6% | not dated | frontierswe.com |
| 21 | zai-org/GLM-5.3-FlashEpoch: glm-5.3-flash_max, Z.ai (Zhipu AI) | open | effort max; Harness: proximus | 18.1% | not dated | frontierswe.com |
| 22 | qwen3.8-max-0902_xhighAlibaba | closed | effort xhigh; Harness: proximus | 17.8% | not dated | frontierswe.com |
| 23 | qwen3.8-max_xhighAlibaba | closed | effort xhigh; Harness: proximus | 15.8% | not dated | frontierswe.com |
| 24 | muse-spark-1.2_xhighMeta AI | closed | effort xhigh; Harness: proximus | 12.0% | not dated | frontierswe.com |
| 25 | thinkingmachines/InklingEpoch: Inkling_xhigh, Thinking Machines | open | effort xhigh; Harness: proximus | 4.1% | not dated | frontierswe.com |
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