BlogLLM Benchmark Scores as Published, Side by Side
ARC-AGI-3: Results as Published
Every result the catalog holds for ARC-AGI-3: 8 models from 3 developers, read from 3 pages. 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 ARC Prize (ARC Prize, Inc.)Introduced 2026Data version 2026103001, as of 2026-10-30
ARC-AGI-3 is the first fully interactive benchmark in the ARC series, made of hundreds of turn-based environments and thousands of game-style levels designed by human game designers. The agent gets no instructions, rules or goals, so it must explore, work out how each environment behaves and what winning means, and carry that understanding into harder levels. It measures whether a system can learn and adapt to unfamiliar situations rather than follow instructions.
ARC-AGI-3
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 AstraOpenAI, released 2026-09-03 | 99.9% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
| GPT-5.6 LunaOpenAI, released 2026-07-09 | 0.18% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 7.78% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 2 more pages | - |
| GPT-5.6 TerraOpenAI, released 2026-07-09 | 0.8% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| GPT-5.5OpenAI, released 2026-04-23 | 0.43% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 30.2% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
| Claude Opus 4.8Anthropic, released 2026-05-28 | 1.5% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
| Gemini 3.1 Pro PreviewGoogle, released 2026-02-19 | 0.42% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 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.