# ARC-AGI-2 leaderboard

Reasoning benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelchoice.dev/leaderboards/benchmarks/reasoning-arc-agi2

Abstraction and Reasoning Corpus for AGI v2. Every model the catalog carries a published ARC-AGI-2 value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Reasoning |
| Measure | Grid transformation puzzles with novel rules |
| Tasks | Visual pattern completion and abstract reasoning |
| Difficulty | Expert-level — hardest public reasoning benchmark |
| Published by | [ARC-AGI-2: A Harder General Intelligence Benchmark](https://arcprize.org/arc-agi/2/) |

ARC-AGI-2 tests whether a system can infer a transformation from example grids and apply it to a new grid. Compare accuracy with the evaluation split, attempt limit, and cost attached.

## ARC-AGI-2 ranking

50 models with a published ARC-AGI-2 value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Grid transformation puzzles with novel rules | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [GPT-6.2 Astra](https://modelchoice.dev/models/gpt-6-2-astra) | OpenAI | 95.833333 | Observed 2026-11-21 · source benchlm:models |
| 2 | [GPT-6 Astra](https://modelchoice.dev/models/gpt-6-astra) | OpenAI | 95 | Observed 2026-11-21 · source benchlm:models |
| 3 | [GPT-6.1 Sol](https://modelchoice.dev/models/gpt-6-1-sol) | OpenAI | 94.166667 | Observed 2026-11-21 · source benchlm:models |
| 4 | [GPT-5.6 Sol](https://modelchoice.dev/models/gpt-5-6-sol) | OpenAI | 92.5 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Gemini 4 Carbon](https://modelchoice.dev/models/gemini-4-carbon) | Google | 92.083333 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Claude Opus 5.5](https://modelchoice.dev/models/claude-opus-5-5) | Anthropic | 91.666667 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Gemini 4 Argon](https://modelchoice.dev/models/gemini-4-argon) | Google | 91.25 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Claude Opus 5](https://modelchoice.dev/models/claude-opus-5) | Anthropic | 90.42 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Claude Fable 5.1](https://modelchoice.dev/models/claude-fable-5-1) | Anthropic | 90 | Observed 2026-11-21 · source benchlm:models |
| 10 | [GPT-6 Sol](https://modelchoice.dev/models/gpt-6-sol) | OpenAI | 89.583333 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Claude Fable 5](https://modelchoice.dev/models/claude-fable-5) | Anthropic | 89.166667 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Gemini 3.8 Flash](https://modelchoice.dev/models/gemini-3-8-flash) | Google | 89.166667 | Observed 2026-11-21 · source benchlm:models |
| 13 | [GPT-6.2 Sol](https://modelchoice.dev/models/gpt-6-2-sol) | OpenAI | 85.833333 | Observed 2026-11-21 · source benchlm:models |
| 14 | [GPT-5.5](https://modelchoice.dev/models/gpt-5-5) | OpenAI | 85 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Gemini 3.7 Flash](https://modelchoice.dev/models/gemini-3-7-flash) | Google | 84.583333 | Observed 2026-11-21 · source benchlm:models |
| 16 | [GPT-5.5 Pro](https://modelchoice.dev/models/gpt-5-5-pro) | OpenAI | 84.2 | Observed 2026-11-21 · source benchlm:models |
| 17 | [GPT-5.4 Pro](https://modelchoice.dev/models/gpt-5-4-pro) | OpenAI | 83.3 | Observed 2026-11-21 · source benchlm:models |
| 18 | [GPT-5.6 Terra](https://modelchoice.dev/models/gpt-5-6-terra) | OpenAI | 82.5 | Observed 2026-11-21 · source benchlm:models |
| 19 | [Inkling 2 Small](https://modelchoice.dev/models/inkling-2-small) | Thinking Machines Lab | 81.666667 | Observed 2026-11-21 · source benchlm:models |
| 20 | [dots3-note Preview](https://modelchoice.dev/models/dots3-note-preview) | Dots Studio | 81.4 | Observed 2026-11-21 · source benchlm:models |
| 21 | [Gemini 3.1 Pro](https://modelchoice.dev/models/gemini-3-1-pro) | Google | 77.08 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Claude Opus 4.7](https://modelchoice.dev/models/claude-opus-4-7) | Anthropic | 75.8 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Claude Opus 4.7 (Adaptive)](https://modelchoice.dev/models/claude-opus-4-7-max) | Anthropic | 75.8 | Last good 2026-11-18 · source benchlm:models |
| 24 | [GPT-5.4](https://modelchoice.dev/models/gpt-5-4) | OpenAI | 73.95 | Observed 2026-11-21 · source benchlm:models |
| 25 | [Gemini 3.5 Flash](https://modelchoice.dev/models/gemini-3-5-flash) | Google | 72.1 | Observed 2026-11-21 · source benchlm:models |
| 26 | [Claude Opus 4.8](https://modelchoice.dev/models/claude-opus-4-8) | Anthropic | 72.08 | Observed 2026-11-21 · source benchlm:models |
| 27 | [Claude Opus 4.6 (Adaptive)](https://modelchoice.dev/models/claude-opus-4-6-thinking) | Anthropic | 68.75 | Observed 2026-11-21 · source benchlm:models |
| 28 | [Grok 4.6](https://modelchoice.dev/models/grok-4-6) | xAI | 67.083333 | Observed 2026-11-21 · source benchlm:models |
| 29 | [DeepSeek V4 Flash 0731](https://modelchoice.dev/models/deepseek-v4-flash-max) | DeepSeek | 61.388889 | Observed 2026-11-21 · source benchlm:models |
| 29 | [Grok 4.7](https://modelchoice.dev/models/grok-4-7) | xAI | 61.388889 | Observed 2026-11-21 · source benchlm:models |
| 31 | [Claude Haiku 5.5](https://modelchoice.dev/models/claude-haiku-5-5) | Anthropic | 61.25 | Observed 2026-11-21 · source benchlm:models |
| 31 | [DeepSeek V4 Pro](https://modelchoice.dev/models/deepseek-v4-pro) | DeepSeek | 61.25 | Observed 2026-11-21 · source benchlm:models |
| 31 | [DeepSeek V4 Pro 0813](https://modelchoice.dev/models/deepseek-v4-pro-max) | DeepSeek | 61.25 | Last good 2026-11-18 · source benchlm:models |
| 34 | [Gemini 3.6 Flash](https://modelchoice.dev/models/gemini-3-6-flash) | Google | 60.416667 | Observed 2026-11-21 · source benchlm:models |
| 34 | [Kimi K3](https://modelchoice.dev/models/kimi-3) | Moonshot AI | 60.416667 | Observed 2026-11-21 · source benchlm:models |
| 36 | [GPT-5.6 Luna](https://modelchoice.dev/models/gpt-5-6-luna) | OpenAI | 59.54 | Observed 2026-11-21 · source benchlm:models |
| 37 | [GPT-6 Luna](https://modelchoice.dev/models/gpt-6-luna) | OpenAI | 59.305556 | Observed 2026-11-21 · source benchlm:models |
| 38 | [Claude Sonnet 4.6](https://modelchoice.dev/models/claude-sonnet-4-6) | Anthropic | 58.33 | Observed 2026-11-21 · source benchlm:models |
| 39 | [Grok 4.20](https://modelchoice.dev/models/grok-4-20-beta) | xAI | 53.3 | Observed 2026-11-21 · source benchlm:models |
| 40 | [GPT-5.2](https://modelchoice.dev/models/gpt-5-2) | OpenAI | 52.9 | Observed 2026-11-21 · source benchlm:models |
| 41 | [Grok 4.5](https://modelchoice.dev/models/grok-4-5) | xAI | 52.64 | Observed 2026-11-21 · source benchlm:models |
| 42 | [Gemini 3 Pro Deep Think](https://modelchoice.dev/models/gemini-3-pro-deep-think) | Google | 45.1 | Last good 2026-11-18 · source benchlm:models |
| 43 | [Muse Spark](https://modelchoice.dev/models/muse-spark) | Meta | 42.5 | Observed 2026-11-21 · source benchlm:models |
| 44 | [Inkling-Small](https://modelchoice.dev/models/inkling-small) | Thinking Machines Lab | 40.138889 | Observed 2026-11-21 · source benchlm:models |
| 45 | [Claude Opus 4.5 Thinking](https://modelchoice.dev/models/claude-opus-4-5-thinking) | Anthropic | 37.64 | Observed 2026-11-21 · source benchlm:models |
| 46 | [Gemini 3 Pro](https://modelchoice.dev/models/gemini-3-pro) | Google | 31.1 | Observed 2026-11-21 · source benchlm:models |
| 47 | [GPT-5.4 mini](https://modelchoice.dev/models/gpt-5-4-mini) | OpenAI | 18.9 | Observed 2026-11-21 · source benchlm:models |
| 48 | [Claude Sonnet 4.5 Thinking](https://modelchoice.dev/models/claude-sonnet-4-5-thinking) | Anthropic | 13.61 | Observed 2026-11-21 · source benchlm:models |
| 49 | [Claude Sonnet 4.5](https://modelchoice.dev/models/claude-sonnet-4-5) | Anthropic | 13.6 | Observed 2026-11-21 · source benchlm:models |
| 50 | [GPT-5.4 nano](https://modelchoice.dev/models/gpt-5-4-nano) | OpenAI | 5.7 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
