Blueprint-Bench 2
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://andonlabs.com/evals/blueprint-bench-2
- Score shown
- Epoch's column "Score", as a percentage, higher is better. The source gives no error bars.
- Results
- 38 in all: 2 for open-weight models, 36 for closed ones. 2 linked to a model page here.
Best open-weight result: moonshotai/Kimi-K3, 29.5%, number 24 of 38 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | gemini-4-carbon_unknownGoogle DeepMind | closed | effort unknown | 58.6% | not dated | andonlabs.com |
| 2 | claude-fable-5-2_unknownAnthropic | closed | effort unknown | 56.6% | not dated | andonlabs.com |
| 3 | gemini-4-argon_unknownGoogle DeepMind | closed | effort unknown | 54.4% | not dated | andonlabs.com |
| 4 | claude-opus-5-5_unknownAnthropic | closed | effort unknown | 51.2% | not dated | andonlabs.com |
| 5 | gpt-6-astra_unknownOpenAI | closed | effort unknown | 49.7% | not dated | andonlabs.com |
| 6 | gpt-6.1-sol_unknownOpenAI | closed | effort unknown | 43.2% | not dated | andonlabs.com |
| 7 | gpt-6.1-terra_unknownOpenAI | closed | effort unknown | 42.1% | not dated | andonlabs.com |
| 8 | claude-fable-5-1_unknownAnthropic | closed | effort unknown | 41.9% | not dated | andonlabs.com |
| 9 | claude-sonnet-5-5_unknownAnthropic | closed | effort unknown | 39.3% | not dated | andonlabs.com |
| 10 | claude-fable-5_unknownAnthropic | closed | effort unknown | 38.6% | not dated | andonlabs.com |
| 11 | gemini-3.8-flash_unknownGoogle DeepMind | closed | effort unknown | 38.6% | not dated | andonlabs.com |
| 12 | gpt-6-sol_unknownOpenAI | closed | effort unknown | 36.9% | not dated | andonlabs.com |
| 13 | gpt-5.5_unknownOpenAI | closed | effort unknown | 36.2% | not dated | andonlabs.com |
| 14 | gemini-3.5-flash_unknownGoogle DeepMind | closed | effort unknown | 33.6% | not dated | andonlabs.com |
| 15 | gpt-5.6-sol_unknownOpenAI | closed | effort unknown | 33.6% | not dated | andonlabs.com |
| 16 | grok-4.6_unknownxAI | closed | effort unknown | 33.2% | not dated | andonlabs.com |
| 17 | claude-haiku-5-5_unknownAnthropic | closed | effort unknown | 33.0% | not dated | andonlabs.com |
| 18 | grok-4.7_unknownxAI | closed | effort unknown | 32.5% | not dated | andonlabs.com |
| 19 | muse-wonder-1.0_unknownMeta AI | closed | effort unknown | 31.7% | not dated | andonlabs.com |
| 20 | gemini-3.6-flash_unknownGoogle DeepMind | closed | effort unknown | 31.2% | not dated | andonlabs.com |
| 21 | gpt-6-luna_unknownOpenAI | closed | effort unknown | 31.2% | not dated | andonlabs.com |
| 22 | gpt-5.6-terra_unknownOpenAI | closed | effort unknown | 30.8% | not dated | andonlabs.com |
| 23 | claude-opus-5_unknownAnthropic | closed | effort unknown | 30.4% | not dated | andonlabs.com |
| 24 | moonshotai/Kimi-K3Epoch: kimi-k3_unknown, Moonshot | open | effort unknown | 29.5% | not dated | andonlabs.com |
| 25 | grok-4.5_unknownxAI | closed | effort unknown | 27.3% | not dated | andonlabs.com |
| 26 | gpt-5.4-2026-03-05_unknownOpenAI | closed | effort unknown | 27.1% | not dated | andonlabs.com |
| 27 | gemini-3.1-pro-previewGoogle DeepMind | closed | default | 26.5% | not dated | andonlabs.com |
| 28 | claude-sonnet-5_unknownAnthropic | closed | effort unknown | 24.9% | not dated | andonlabs.com |
| 29 | claude-opus-4-7_unknownAnthropic | closed | effort unknown | 24.5% | not dated | andonlabs.com |
| 30 | gpt-5.6-luna_unknownOpenAI | closed | effort unknown | 22.6% | not dated | andonlabs.com |
| 31 | claude-opus-4-8_unknownAnthropic | closed | effort unknown | 14.5% | not dated | andonlabs.com |
| 32 | claude-sonnet-4-6_unknownAnthropic | closed | effort unknown | 6.7% | not dated | andonlabs.com |
| 33 | moonshotai/Kimi-K2.6Epoch: kimi-k2.6, Moonshot | open | default | 3.9% | not dated | andonlabs.com |
| 34 | claude-haiku-4-5-20251001_unknownAnthropic | closed | effort unknown | 0.0% | not dated | andonlabs.com |
| 35 | gemini-3-flash-previewGoogle DeepMind | closed | default | 0.0% | not dated | andonlabs.com |
| 36 | gemini-robotics-er-1.6-previewGoogle | closed | default | 0.0% | not dated | andonlabs.com |
| 37 | grok-4-20xAI | closed | default | 0.0% | not dated | andonlabs.com |
| 38 | grok-4.3_unknownxAI | closed | effort unknown | 0.0% | not dated | andonlabs.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