BlogLLM Benchmark Scores as Published, Side by Side
ZebraLogic: Results as Published
Every result the catalog holds for ZebraLogic: 7 models from 6 developers, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area General Knowledge and ReasoningMaintainer Bill Yuchen Lin and co-authors (University of Washington, Allen Institute for AI, and Stanford University)Introduced 2025Data version 2026103001, as of 2026-10-30
ZebraLogic tests logical reasoning on logic grid puzzles, also called zebra puzzles, built as constraint satisfaction problems. The benchmark has 1,000 puzzles whose size and clue complexity can be controlled. Each puzzle has houses and attributes, and the model must give each house a unique value for every attribute so that every clue holds. Overall accuracy is the share of puzzles solved completely, and cell accuracy is the share of individual cells filled correctly.
ZebraLogic
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-4.1OpenAI, released 2025-04-14 | 58.5% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| Claude Opus 4Anthropic, released 2025-05-22 | 59.3% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: no extended thinking |
| Claude Sonnet 4Anthropic, released 2025-05-22 | 79.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: no extended thinking |
| Gemini 2.5 FlashGoogle, released 2025-05-20 | 57.9% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| DeepSeek-V3-0324DeepSeek, released 2025-03-24 | 84% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| Qwen3-235B-A22BAlibaba, released 2025-04-29 | 37.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: non-thinking |
| Kimi K2 InstructMoonshot AI, released 2025-07-11 | 89% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.