Reasoning

Epoch live leaderboard

ARC-AGI-2

Abstract reasoning and pattern generalization on grid-based tasks.

Published source

Metric

Accuracy

Results shown

235

Publisher

Epoch AI live leaderboard

Snapshot

Fetched 2026-10-08T23:19:28.137Z

235 of 235 models

Higher is better

GPT-6 Astra (max)

GPT-6.1 Sol (max)

GPT-6 Astra (xhigh)

Claude Opus 5.5 (high)

GPT-5.6 Sol (max)

Claude Opus 5.5 (xhigh)

GPT-6 Astra (high)

GPT-6 Astra (medium)

Claude Opus 5.5 (max)

GPT-6.1 Sol (xhigh)

GPT-6.1 Sol (high)

Claude Opus 5 (max)

GPT-5.6 Sol (xhigh)

Claude Fable 5.1 (max)

Claude Fable 5.1 (xhigh)

GPT-6 Sol (max)

Claude Fable 5 (max)

Gemini 3.8 Flash (high)

Claude Fable 5.1 (high)

Claude Fable 5 (xhigh)

Claude Opus 5 (high)

Claude Fable 5 (high)

Claude Opus 5.5 (medium)

GPT-6.1 Sol (medium)

Claude Fable 5.1 (medium)

GPT-5.6 Sol (high)

GPT-6 Astra (low)

GPT-5.5 (xhigh)

Gemini 3.7 Flash (high)

Gemini 3 Deep Think Preview

GPT-5.5 Pro (high)

GPT-5.5 Pro (xhigh)

GPT-5.6 Terra (max)

GPT-5.4 Pro (xhigh)

GPT-5.5 (high)

Gemini 3.8 Flash (medium)

Claude Fable 5 (medium)

Claude Fable 5.1 (low)

GPT-6 Sol (xhigh)

Gemini 3.8 Flash (low)

Gemini 3.1 Pro Preview

Claude Fable 5 (low)

GPT-6.1 Sol (low)

Claude Opus 4.7 (max)

GPT-5.6 Terra (xhigh)

GPT-5.4 (xhigh)

Gemini 3.5 Flash (high)

Claude Opus 4.8 (high)

Claude Opus 4.8 (medium)

GPT-5.5 (medium)

Claude Opus 5.5 (low)

Claude Opus 4.6 (120k thinking)

GPT-6 Sol (high)

Claude Opus 4.7 (high)

GPT-5.4 (high)

Claude Opus 4.7 (medium)

Grok 4.6 (xhigh)

GPT-5.6 Sol (medium)

GPT-5.6 Terra (high)

GLM-5.3-Flash (max)

Grok 4.20

Grok 4.6 (high)

Gemini 3.7 Flash (medium)

Claude Opus 4.8 (low)

Claude Opus 4.7 (low)

DeepSeek V4 Flash 0731 (max)

Grok 4.6 (medium)

DeepSeek V4 Pro 0813 (max)

Claude Sonnet 4.6 (high)

Kimi K3 (Max)

Gemini 3.6 Flash (high)

DeepSeek V4 Pro 0813 (high)

GPT-6 Astra (none)

GPT-5.6 Luna (max)

GPT-6 Luna (max)

Claude Sonnet 4.6 (max)

GPT-6 Sol (medium)

DeepSeek V4 Pro 0813 (low)

DeepSeek V4 Flash 0731 (high)

GPT-5.4 (medium)

Kimi K3 (high)

GPT-5.2 Pro (high)

Gemini 3.7 Flash (low)

GPT-5.2 (xhigh)

Grok 4.5 (high)

Grok 4.5 (medium)

Gemini 3.6 Flash (medium)

GLM-5.3-Flash (high)

GPT-5.6 Luna (xhigh)

DeepSeek V4 Flash 0731 (low)

GPT-5.2 (high)

GPT-5.6 Sol (low)

Qwen3.8 27B (xhigh)

GPT-6 Luna (xhigh)

Inkling Small (xhigh)

GPT-5.2 Pro (medium)

Claude Opus 4.5 (64k thinking)

GPT-5.6 Terra (medium)

Inkling

Gemini 3 Flash Preview

GPT-5.5 (low)

Grok 4.5 (low)

Inkling Small (high)

GPT-6 Sol (low)

GPT-6 Luna (high)

Gemini 3 Pro Preview

Gemini 3.6 Flash (low)

GPT-5.6 Luna (high)

GPT-5.4 (low)

GLM-5.3-Flash (low)

Grok 4.6 (low)

GPT-5.2 (medium)

Claude Opus 4.5 (16k thinking)

Qwen3.8 27B (low)

GLM-5.2 (unknown thinking)

GPT-5.4 mini (xhigh)

GPT-5.6 Terra (low)

GPT-5 Pro (unknown thinking)

GPT-6 Luna (medium)

GPT-5.1 (high)

Grok 4

Claude Opus 4.5 (8k thinking)

Inkling Small (medium)

Claude Sonnet 4.5 (32k thinking)

Qwen3.8 27B (medium)

GPT-5.4 mini (high)

Kimi K3 (low)

Kimi K2.5

Gemini 3.5 Flash-Lite (high)

GPT-5 (high)

GPT-5.2 (low)

Gemini 3.5 Flash (minimal)

Claude Opus 4 (16k thinking)

Claude Opus 4.5 (no thinking)

GPT-5 (medium)

GPT-5.6 Luna (medium)

Claude Sonnet 4.5 (16k thinking)

Claude Sonnet 4.5 (2k thinking)

o3 (high)

GPT-5.1 (medium)

tiny-recursion-model

o4-mini (high)

Claude Sonnet 4 (16k thinking)

Claude Sonnet 4.5 (2k thinking)

GPT-5.4 nano (xhigh)

Grok 4 Fast

Gemini 3.5 Flash-Lite (medium)

GPT-5.6 Luna (low)

Inkling Small (low)

o3-pro

Gemini 2.5 Pro (32k thinking)

MiniMax-M2.5

GLM-5

GPT-6 Luna (low)

Claude Opus 4 (8k thinking)

GPT-5 mini (high)

GPT-5.4 mini (medium)

Gemini 2.5 Pro (16k thinking)

GPT-5 mini (medium)

Claude Haiku 4.5 (32k thinking)

DeepSeek-V3.2 (Thinking; Novita)

Inkling Small (minimal)

Claude Sonnet 4.5 (no thinking)

GPT-5.4 nano (high)

o3-mini (high)

o3 (medium)

Gemini 2.5 Pro (8k thinking)

Claude Haiku 4.5 (16k thinking)

Gemini 3.6 Flash (minimal)

GPT-5 nano (high)

Gemini 2.5 Flash

o4-mini (medium)

Gemini 2.5 Flash

Claude Sonnet 4 (8k thinking)

DeepSeek V4 Flash 0731 (none)

o3-mini (medium)

o3-pro

o3 (low)

Gemini 2.5 Flash

o3-pro

GPT-5 (low)

GPT-5.1 (low)

GPT-5.4 nano (medium)

Gemini 2.5 Flash

o4-mini (low)

GPT-5 mini (medium)

Claude Haiku 4.5 (8k thinking)

GPT-6 Sol (none)

GPT-5.4 nano (low)

Gemini 3.5 Flash-Lite (low)

Qwen3.8 27B (no thinking)

Gemini 2.0 Flash (Feb 2025)

DeepSeek-R1

Claude Sonnet 4

codex-mini-2025-05-16

Claude Opus 4

Qwen3 Non-thinking (Jul 2025)

Claude Haiku 4.5

Claude Haiku 4.5 (1k thinking)

DeepSeek-R1 (May 2025)

GPT-5.4 mini (low)

Claude 3.7 Sonnet (8k thinking)

GPT-5 nano (medium)

Claude Sonnet 4 (1k thinking)

Gemini 3.5 Flash-Lite (minimal)

DeepSeek V4 Pro 0813 (none)

o1-mini (unknown thinking)

GPT-5.2 (unknown thinking)

GPT-5 mini (medium)

GPT-4.5 Preview (Feb 2025)

Gemini 1.5 Pro

Claude 3.7 Sonnet (16k thinking)

GPT-4.1

Grok 3 mini (low)

GPT-5.1 (no thinking)

Claude 3.7 Sonnet (1k thinking)

Claude 3.7 Sonnet

GPT-4o mini

GPT-4.1 nano

Llama 4 Scout

Llama 4 Maverick

o3-mini

GPT-4.1 mini

Grok 3

Claude Opus 4 (1k thinking)

Magistral Small 1.0

Magistral Medium 1.1

Gemini 2.5 Pro Preview (Jun 2025, 1k thinking)

GPT-5 (minimal)

GPT-5 nano (low)

GPT-4o (Nov 2024)

GPT-5 nano (low)

Inkling Small (none)

GPT-6 Luna (none)

List bars visualize relative performance. Plot bars use published 0–100 scores, shrink to a 1.5rem minimum, then scroll horizontally.

Scores are published snapshots and are only comparable within the same benchmark, version, metric, and evaluation configuration.