Which model leads on ARC-AGI?

Claude Fable 5 from Anthropic leads ARC-AGI with a score of 98.5. The median score across 79 tested models is 72.8.

Is ARC-AGI saturated?

Yes · the top model on ARC-AGI has reached 98.5 out of 100, within 5% of the theoretical ceiling. This benchmark is approaching saturation and may be replaced by a harder successor.

Does ARC-AGI predict performance on other benchmarks?

Yes · ARC-AGI scores correlate 1.00 with Artificial Analysis · Terminal-Bench Hard across 5 shared models. Models that do well on ARC-AGI tend to do well on Artificial Analysis · Terminal-Bench Hard.

How often is ARC-AGI data refreshed?

BenchGecko pulls updates daily. New model scores on ARC-AGI appear as soon as they are published by Epoch AI or the model provider.

Benchmark · ReasoningSettled

ARC-AGI

Name: ARC-AGI Benchmark
Creator: BenchGecko
License: https://creativecommons.org/licenses/by/4.0/

ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.

Updated 2026-09-29

Models tested

Top score

98.5

Claude Fable 5

Median

72.8

min 0.5

Top-5 spread

σ 0.0

Settled

The Frontier

Best score over time · one chart, every benchmark

Chart type

Frontier on ARC-AGI rose from 4.5 to 98.5 in 19 months · +94.0 points · latest leader Claude Fable 5 from Anthropic.

Pink dots = frontier records · 13 totalClick to open model page

Full rankings

79 models tested · sorted by score

#	Model	Score	Price
1	Claude Fable 5· Anthropic	98.5	$10.00
2	Claude Opus 5.5· Anthropic	98.5	$4.00
3	Gemini 3.8 Flash· Google DeepMind	98.5	$0.75
4	GPT-6 Astra· OpenAI	98.5	$10.00
5	GPT-6.1 Sol· OpenAI	98.5	$2.00
6	Gemini 3.1 Pro Preview· Google DeepMind	98.0	$2.00
7	Claude Fable 5.1· Anthropic	97.5	$10.00
8	Claude Opus 5· Anthropic	97.5	$5.00
9	GPT-5.6 Sol· OpenAI	97.5	$2.00
10	GPT-5.6 Terra· OpenAI	96.5	$2.00
11	Gemini 3.7 Flash· Google DeepMind	95.5	$0.75
12	GPT-6 Sol· OpenAI	95.5	$2.00
13	GPT-5.5· OpenAI	95.0	$5.00
14	DeepSeek V4.1 Flash· DeepSeek	94.5	$0.15
15	GPT-5.4 Pro· OpenAI	94.5	$30.00
16	Kimi K3· moonshotai	94.5	$0.45
17	Claude Opus 4.6· Anthropic	94.0	$5.00
18	GPT-5.4· OpenAI	93.7	$2.50
19	Claude Opus 4.7· Anthropic	93.5	$5.00
20	Claude Opus 4.8· Anthropic	92.5	$5.00
21	Gemini 3.5 Flash· Google DeepMind	92.5	$1.50
22	Gemini 3.6 Flash· Google DeepMind	91.2	$0.75
23	GLM 5.3 Flash· z-ai	91.0	$0.15
24	DeepSeek V4 Pro 0813· DeepSeek	90.5	$0.66
25	GPT-5.2 Pro· OpenAI	90.5	$21.00
26	Grok 4.7· xAI	90.2	$2.00
27	Grok 4.20· xAI	89.5	$1.25
28	DeepSeek V4 Flash 0731· DeepSeek	89.0	$0.02
29	GPT-5.6 Luna· OpenAI	88.0	$0.20
30	Grok 4.6· xAI	87.5	$2.00
31	Qwen3.8 27B· Alibaba Qwen	87.5	$0.42
32	Grok 4.5· xAI	87.2	$2.00
33	GPT-6 Luna· OpenAI	86.7	$0.10
34	Claude Sonnet 4.6· Anthropic	86.5	$3.00
35	GPT-5.2· OpenAI	86.2	$1.75
36	Gemini 3 Flash Preview· Google DeepMind	84.7	$0.50
37	Claude Opus 4.5· Anthropic	80.0	$5.00
38	GLM 5.2· z-ai	77.0	$0.02
39	Gemini 3 Pro· Google DeepMind	75.0	n/a
40	GPT-5.1· OpenAI	72.8	$1.25
41	GPT-5 Pro· OpenAI	70.2	$15.00
42	Grok 4· xAI	66.7	n/a
43	GPT-5· OpenAI	65.7	$1.25
44	Kimi K2.5· moonshotai	65.3	$0.45
45	Claude Sonnet 4.5· Anthropic	63.7	$3.00
46	GPT-5.4 Mini· OpenAI	63.7	$0.75
47	MiniMax M2.5· minimax	63.7	$0.27
48	o3· OpenAI	60.8	$2.00
49	o3 Pro· OpenAI	59.3	$20.00
50	o4 Mini· OpenAI	58.7	$1.10
51	DeepSeek V3.2· DeepSeek	57.0	$0.28
52	GPT-5 Mini· OpenAI	54.3	$0.25
53	Gemini 3.5 Flash Lite· Google DeepMind	53.5	$0.30
54	GPT-5.4 Nano· OpenAI	51.5	$0.20
55	Grok 4 Fast· xAI	48.5	n/a
56	Claude Haiku 4.5· Anthropic	47.7	$1.00
57	GLM 5· z-ai	44.7	$1.00
58	Gemini 2.5 Pro· Google DeepMind	41.0	$1.25
59	Claude Sonnet 4· Anthropic	40.0	$3.00
60	Claude Opus 4· Anthropic	35.7	$15.00
61	o3 Mini· OpenAI	34.5	$1.10
62	Gemini 2.5 Flash· Google DeepMind	32.3	$0.30
63	o1· OpenAI	30.7	$15.00
64	Claude 3.7 Sonnet· Anthropic	28.6	n/a
65	R1 0528· DeepSeek	21.2	$0.50
66	GPT-5 Nano· OpenAI	20.7	$0.05
67	o1-preview· OpenAI	18.0	n/a
68	Grok-3 mini· xAI	16.5	n/a
69	R1· DeepSeek	15.8	$0.70
70	o1-mini· OpenAI	14.0	n/a
71	Qwen3 235B A22B Instruct 2507· Alibaba Qwen	11.0	$0.09
72	GPT-4.5· OpenAI	10.3	n/a
73	GPT-4.1· OpenAI	5.5	$2.00
74	Grok 3· xAI	5.5	n/a
75	U Magistral Small 1.1· Unknown	5.0	n/a
76	GPT-4o (2024-11-20)· OpenAI	4.5	$2.50
77	Llama 4 Maverick· Meta	4.4	$0.19
78	GPT-4.1 Mini· OpenAI	3.5	$0.40
79	Llama 4 Scout· Meta	0.5	$0.10