Last verified
ReasoningOfficial + Vendor scoresPrices verified

ARC-AGI-1 leaderboard

As of November 8, 2026, 6 systems share the top ARC-AGI-1 score, 98.5%: Claude Fable 5, Claude Opus 5.5, Gemini 3.8 Flash, Gemini 4 Argon and 2 more. Next is Gemini 3.1 Pro Preview at 98%. 99 systems ranked. All 99 scores on the main board come from ARC Prize Foundation, not from the labs that make the models.

See the ranking →All benchmarks →Model prices →
Top score
98.5%
6-way tie
Systems ranked
99
one setting
Independent rows
99/101
official leaderboard
Cheapest near top
$0.15/1M
DeepSeek V4.1 Flash · within 5 pts
With API prices
69/99
rates from our price index
Newest score
Nov 5
2026 · newest result

What is ARC-AGI-1? ARC-AGI-1 is the original Abstraction and Reasoning Corpus: grid puzzles where a model must infer a hidden rule from a few examples and apply it to a new input. Made by François Chollet; maintained by the ARC Prize Foundation, 2019.

§ 01 / LEADERBOARD

Top models on ARC-AGI-1, November 2026

99 entries
# Model Score Price / 1M Context Source
1 Claude Fable 5 Anthropic
98.5%
ARC Prize verified, best result 		$10 / $50 1M Official leaderboard ARC Prize Foundation
1 Claude Opus 5.5 Anthropic
98.5%
ARC Prize verified, best result 		$4 / $20 1M Official leaderboard ARC Prize Foundation
1 Gemini 3.8 Flash Google DeepMind
98.5%
ARC Prize verified, best result 		$0.75 / $3.75 1M Official leaderboard ARC Prize Foundation
1 Gemini 4 Argon Google DeepMind
98.5%
ARC Prize verified, best result 		$2 / $10 1M Official leaderboard ARC Prize Foundation
1 GPT-6 Astra OpenAI
98.5%
ARC Prize verified, best result 		$10 / $50 1.05M Official leaderboard ARC Prize Foundation
1 GPT-6.1 Sol OpenAI
98.5%
ARC Prize verified, best result 		$2 / $10 1.05M Official leaderboard ARC Prize Foundation
7 Gemini 3.1 Pro Preview Preview Google
98%
ARC Prize verified, best result 		$2 / $12 1M Official leaderboard ARC Prize Foundation
8 Claude Fable 5.1 Anthropic
97.5%
ARC Prize verified, best result 		$10 / $50 1M Official leaderboard ARC Prize Foundation
8 Claude Opus 5 Anthropic
97.5%
ARC Prize verified, best result 		$5 / $25 1M Official leaderboard ARC Prize Foundation
8 GPT-5.6 Sol OpenAI
97.5%
ARC Prize verified, best result 		$4 / $20 1.05M Official leaderboard ARC Prize Foundation
11 Claude Sonnet 5.5 Anthropic
96.5%
ARC Prize verified, best result 		$2 / $10 1M Official leaderboard ARC Prize Foundation
11 GPT-5.5 Pro OpenAI
96.5%
ARC Prize verified, best result 		$30 / $180 1.05M Official leaderboard ARC Prize Foundation
11 GPT-5.6 Terra OpenAI
96.5%
ARC Prize verified, best result 		$2 / $12 1.05M Official leaderboard ARC Prize Foundation
11 Inkling 2 Small Not in our price index
96.5%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
15 Gemini 3 Deep Think (2/26) Not in our price index
96%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
16 Gemini 3.7 Flash Deprecated Google DeepMind
95.5%
ARC Prize verified, best result 		$0.75 / $3.75 1M Official leaderboard ARC Prize Foundation
16 GPT-6 Sol OpenAI
95.5%
ARC Prize verified, best result 		$2 / $10 1.05M Official leaderboard ARC Prize Foundation
18 GPT-5.5 OpenAI
95%
ARC Prize verified, best result 		$5 / $30 1M Official leaderboard ARC Prize Foundation
19 DeepSeek V4.1 Flash DeepSeek · listed as “DeepSeek V4.1 Flash 1017”
94.5%
ARC Prize verified, best result 		$0.3 / $1.2 1M Official leaderboard ARC Prize Foundation
19 GPT-5.2 (Refinement) Not in our price index
94.5%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
19 GPT-5.4 Pro OpenAI
94.5%
ARC Prize verified, best result 		$30 / $180 1.05M Official leaderboard ARC Prize Foundation
19 Kimi K3 Moonshot (Kimi)
94.5%
ARC Prize verified, best result 		$3 / $15 1M Official leaderboard ARC Prize Foundation
23 Claude Opus 4.6 Anthropic · listed as “Claude Opus 4.6 (120K)”
94%
ARC Prize verified, best result 		$5 / $25 1M Official leaderboard ARC Prize Foundation
24 GPT-5.4 OpenAI
93.7%
ARC Prize verified, best result 		$2.5 / $15 1.05M Official leaderboard ARC Prize Foundation
25 Claude Opus 4.7 Anthropic · listed as “Claude 4.7”
93.5%
ARC Prize verified, best result 		$5 / $25 1M Official leaderboard ARC Prize Foundation
26 Claude Opus 4.8 Anthropic
92.5%
ARC Prize verified, best result 		$5 / $25 1M Official leaderboard ARC Prize Foundation
26 Gemini 3.5 Flash Deprecated Google
92.5%
ARC Prize verified, best result 		$1.5 / $9 1M Official leaderboard ARC Prize Foundation
28 Gemini 3.6 Flash Google
91.2%
ARC Prize verified, best result 		$0.75 / $3.75 1M Official leaderboard ARC Prize Foundation
29 GLM-5.3-Flash Zhipu (Z.ai / GLM)
91%
ARC Prize verified, best result 		$0.15 / $0.5 1M Official leaderboard ARC Prize Foundation
30 GPT-5.6 Luna OpenAI · listed as “GPT-5.6 Luna 2026-07-30”
90.7%
ARC Prize verified, best result 		$0.2 / $1.2 1.05M Official leaderboard ARC Prize Foundation
31 DeepSeek V4 Pro DeepSeek · listed as “DeepSeek V4 Pro 0813”
90.5%
ARC Prize verified, best result 		$1.32 / $3.96 1M Official leaderboard ARC Prize Foundation
31 GPT-5.2 Pro Not in our price index
90.5%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
33 Grok 4.7 SpaceXSI
90.2%
ARC Prize verified, best result 		$2 / $6 500K Official leaderboard ARC Prize Foundation
34 Grok 4.20 (0309) Reasoning SpaceXSI · listed as “Grok 4.20 (Reasoning)”
89.5%
ARC Prize verified, best result 		$1.25 / $2.5 1M Official leaderboard ARC Prize Foundation
35 Claude Haiku 5.5 Anthropic
89%
ARC Prize verified, best result 		$0.1 / $0.5 1M Official leaderboard ARC Prize Foundation
35 DeepSeek V4 Flash Deprecated DeepSeek · listed as “DeepSeek V4 Flash 0731”
89%
ARC Prize verified, best result 		$0.44 / $1.32 1M Official leaderboard ARC Prize Foundation
35 Muse Spark Open Not in our price index
89%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
38 Gemini 3 Deep Think (Preview) Not in our price index
87.5%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
38 Grok 4.6 SpaceXSI
87.5%
ARC Prize verified, best result 		$2 / $6 500K Official leaderboard ARC Prize Foundation
38 Qwen3.8-27B Alibaba (Qwen)
87.5%
ARC Prize verified, best result 		$0.5 / $3 1M Official leaderboard ARC Prize Foundation
41 Grok 4.5 SpaceXSI
87.2%
ARC Prize verified, best result 		$2 / $6 500K Official leaderboard ARC Prize Foundation
42 GPT-6 Luna OpenAI
86.7%
ARC Prize verified, best result 		$0.1 / $0.5 1.05M Official leaderboard ARC Prize Foundation
43 Claude Sonnet 4.6 Anthropic
86.5%
ARC Prize verified, best result 		$3 / $15 1M Official leaderboard ARC Prize Foundation
44 GPT-5.2 Not in our price index
86.2%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
45 Gemini 3 Flash Preview Preview Google
84.7%
ARC Prize verified, best result 		$0.5 / $3 1M Official leaderboard ARC Prize Foundation
46 Inkling Small Not in our price index
84%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
47 Claude Opus 4.5 Anthropic · listed as “Opus 4.5”
80%
ARC Prize verified, best result 		$5 / $25 200K Official leaderboard ARC Prize Foundation
48 Grok 4 (Refinement) Not in our price index
79.6%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
49 Inkling Not in our price index
79.5%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
50 GLM-5.2 Zhipu (Z.ai / GLM)
77%
ARC Prize verified, best result 		$1.4 / $4.4 1M Official leaderboard ARC Prize Foundation
51 o3 (Preview) Not in our price index
75.7%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
52 Gemini 3 Pro Preview Retired Google · listed as “Gemini 3 Pro”
75%
ARC Prize verified, best result 		$2 / $12 1M Official leaderboard ARC Prize Foundation
53 Beam 1.0 Not in our price index
74%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
54 GPT-5.1 Not in our price index
72.8%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
55 GPT-5 Pro Not in our price index
70.2%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
56 Grok 4 Retired SpaceXSI · listed as “Grok 4 (Thinking)”
66.7%
ARC Prize verified, best result 		$3 / $15 256K Official leaderboard ARC Prize Foundation
57 GPT-5 Not in our price index
65.7%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
58 Kimi K2.5 Retired Moonshot (Kimi)
65.3%
ARC Prize verified, best result 		$0.6 / $3 262K Official leaderboard ARC Prize Foundation
59 Claude Sonnet 4.5 Deprecated Anthropic
63.7%
ARC Prize verified, best result 		$3 / $15 200K Official leaderboard ARC Prize Foundation
59 GPT-5.4 mini OpenAI
63.7%
ARC Prize verified, best result 		$0.75 / $4.5 400K Official leaderboard ARC Prize Foundation
59 MiniMax M2.5 Deprecated MiniMax
63.7%
ARC Prize verified, best result 		$0.3 / $1.2 205K Official leaderboard ARC Prize Foundation
62 o3 Not in our price index
60.8%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
63 o3-Pro Not in our price index
59.3%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
64 o4-mini Retired OpenAI
58.7%
ARC Prize verified, best result 		$1.1 / $4.4 200K Official leaderboard ARC Prize Foundation
65 Deepseek V3.2 Not in our price index
57%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
66 ARChitects Not in our price index
56%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
67 GPT-5 Mini Not in our price index
54.3%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
68 Gemini 3.5 Flash-Lite Google
53.5%
ARC Prize verified, best result 		$0.3 / $2.5 1M Official leaderboard ARC Prize Foundation
69 GPT-5.4 nano Deprecated OpenAI
51.5%
ARC Prize verified, best result 		$0.2 / $1.25 400K Official leaderboard ARC Prize Foundation
70 Grok 4 Fast Retired SpaceXSI · listed as “Grok 4 (Fast Reasoning)”
48.5%
ARC Prize verified, best result 		$0.2 / $0.5 2M Official leaderboard ARC Prize Foundation
71 Claude Haiku 4.5 Anthropic
47.7%
ARC Prize verified, best result 		$1 / $5 200K Official leaderboard ARC Prize Foundation
72 GLM-5 Zhipu (Z.ai / GLM)
44.7%
ARC Prize verified, best result 		$1 / $3.2 200K Official leaderboard ARC Prize Foundation
73 Gemini 2.5 Pro Deprecated Google
41%
ARC Prize verified, best result 		$1.25 / $10 1M Official leaderboard ARC Prize Foundation
74 Claude Sonnet 4 Retired Anthropic
40%
ARC Prize verified, best result 		$3 / $15 200K Official leaderboard ARC Prize Foundation
74 Tiny Recursion Model (TRM) Not in our price index
40%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
76 Claude Opus 4 Retired Anthropic
35.7%
ARC Prize verified, best result 		$15 / $75 200K Official leaderboard ARC Prize Foundation
77 o3-mini Retired OpenAI
34.5%
ARC Prize verified, best result 		$1.1 / $4.4 200K Official leaderboard ARC Prize Foundation
78 Gemini 2.5 Flash Deprecated Google · listed as “Gemini 2.5 Flash (Preview)”
33.3%
ARC Prize verified, best result 		$0.3 / $2.5 1M Official leaderboard ARC Prize Foundation
79 Hierarchical Reasoning Model (HRM) Not in our price index
32%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
80 Claude Sonnet 3.7 Retired Anthropic · listed as “Claude 3.7”
28.6%
ARC Prize verified, best result 		$3 / $15 200K Official leaderboard ARC Prize Foundation
81 Codex Mini (Latest) Not in our price index
27.3%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
82 Nemotron 3.5 Super VL 120B A12B NVIDIA · listed as “Nemotron 3.5 Super VL”
22.7%
ARC Prize verified, best result 		— 1M Official leaderboard ARC Prize Foundation
83 Deepseek R1 (05/28) Not in our price index
21.2%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
84 GPT-5 Nano Not in our price index
20.7%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
85 Icecuber Not in our price index
17%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
86 Grok 3 Mini Retired SpaceXSI
16.5%
ARC Prize verified, best result 		$0.3 / $0.5 131K Official leaderboard ARC Prize Foundation
87 DeepSeek R1 Retired DeepSeek
15.8%
ARC Prize verified, best result 		$0.55 / $2.19 64K Official leaderboard ARC Prize Foundation
88 o1-mini Retired OpenAI
14%
ARC Prize verified, best result 		$3 / $12 128K Official leaderboard ARC Prize Foundation
89 Qwen3 235B A22B Instruct 2507 Retired Alibaba (Qwen)
11%
ARC Prize verified, best result 		$0.23 / $0.92 262K Official leaderboard ARC Prize Foundation
90 GPT-4.5 preview Retired OpenAI · listed as “GPT-4.5”
10.3%
ARC Prize verified, best result 		$75 / $150 128K Official leaderboard ARC Prize Foundation
91 Magistral Medium Retired Mistral
5.9%
ARC Prize verified, best result 		$2 / $5 — Official leaderboard ARC Prize Foundation
92 GPT-4.1 Deprecated OpenAI
5.5%
ARC Prize verified, best result 		$2 / $8 1M Official leaderboard ARC Prize Foundation
92 Grok 3 Retired SpaceXSI
5.5%
ARC Prize verified, best result 		$3 / $15 131K Official leaderboard ARC Prize Foundation
94 Magistral Small Retired Mistral
5%
ARC Prize verified, best result 		$0.5 / $1.5 — Official leaderboard ARC Prize Foundation
95 GPT-4o Deprecated OpenAI
4.5%
ARC Prize verified, best result 		$2.5 / $10 128K Official leaderboard ARC Prize Foundation
96 Llama 4 Maverick Not in our price index
4.4%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
97 GPT-4.1 mini Deprecated OpenAI
3.5%
ARC Prize verified, best result 		$0.4 / $1.6 1M Official leaderboard ARC Prize Foundation
98 Llama 4 Scout Not in our price index
0.5%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
99 GPT-4.1-Nano Not in our price index
0%
ARC Prize verified, best result 		— — Official leaderboard ARC Prize Foundation
§ 02 / SCORE VS PRICE

ARC-AGI-1: score against price

30% 50% 70% 90% $0.1 $0.3 $1 $3 $10 $30 Effective price per 1M tokens (log scale) Score GPT-6 Luna: 86.7% at $0.0682GPT-5.6 Luna: 90.7% at $0.152Gemini 3.5 Flash-Lite: 53.5% at $0.272Gemini 3 Flash Preview: 84.7% at $0.361Gemini 3.6 Flash: 91.2% at $0.481GPT-5.4 mini: 63.7% at $0.541Grok 4.20 (0309) Reasoning: 89.5% at $0.558DeepSeek V4 Pro: 90.5% at $0.569GLM-5: 44.7% at $0.572Claude Haiku 4.5: 47.7% at $0.682Qwen3.8-27B: 87.5% at $0.7GLM-5.2: 77% at $0.78Grok 4.5: 87.2% at $1.04Grok 4.7: 90.2% at $1.19Grok 4.6: 87.5% at $1.19Gemini 4 Argon: 98.5% at $1.21GPT-6.1 Sol: 98.5% at $1.29Claude Sonnet 5.5: 96.5% at $1.29GPT-6 Sol: 95.5% at $1.36Gemini 3.1 Pro Preview: 98% at $1.44GPT-5.6 Terra: 96.5% at $1.52GPT-5.4: 93.7% at $1.8Kimi K3: 94.5% at $1.92Claude Sonnet 4.6: 86.5% at $2.05Claude Opus 5.5: 98.5% at $2.58GPT-5.6 Sol: 97.5% at $2.73Claude Opus 5: 97.5% at $3.41Claude Opus 4.6: 94% at $3.41Claude Opus 4.7: 93.5% at $3.41Claude Opus 4.8: 92.5% at $3.41Claude Opus 4.5: 80% at $3.41GPT-5.5: 95% at $3.61Claude Fable 5.1: 97.5% at $6.26Claude Fable 5: 98.5% at $6.82GPT-6 Astra: 98.5% at $6.82GPT-5.5 Pro: 96.5% at $42GPT-5.4 Pro: 94.5% at $42 Claude Haiku 5.5: 89% at $0.0682 Claude Haiku 5.5 GLM-5.3-Flash: 91% at $0.0875 GLM-5.3-Flash DeepSeek V4.1 Flash: 94.5% at $0.15 DeepSeek V4.1 Flash Gemini 3.8 Flash: 98.5% at $0.481 Gemini 3.8 Flash
The dashed line joins the models that no cheaper model beats: Claude Haiku 5.5 (89%, $0.0682 per 1M tokens), GLM-5.3-Flash (91%, $0.0875 per 1M tokens), DeepSeek V4.1 Flash (94.5%, $0.15 per 1M tokens) and Gemini 3.8 Flash (98.5%, $0.481 per 1M tokens). Anything below the line costs more for a lower score. Price is the effective rate per 1M tokens on current list prices: 92% input, 8% output, 82% of input served from cache.
§ 03 / PROVENANCE

Can you trust ARC-AGI-1 scores?

99 scores on this page come from the benchmark's own leaderboard (ARC Prize Foundation), where every model runs on the same harness. The other 2 are labs' reports on their own models. 2 models have numbers from two sources for the same setting, and every pair agrees within 1 point.

§ 04 / HOW IT WORKS

How ARC-AGI-1 works

The task

A few input and output grid pairs, usually about three, demonstrate a transformation; the model has to produce the output grid for a new test input.

The score

The share of evaluation tasks solved. ARC Prize publishes the cost per task beside every verified score.

Test set

400 public evaluation tasks; 100-task semi-private and private sets, as published by the authors.

Made by

François Chollet; maintained by the ARC Prize Foundation, first published 2019.

Official sources

§ 05 / SOURCES

Sources and freshness

101 published scores from 2 sources. Each was read on its source page; the oldest reading on this page is from November 8, 2026. Prices beside each score come from our model price index and move when a vendor reprices.

ARC-AGI-1 FAQ

Answers derived from the 101 published scores on this page and the current price index.

Q · 01

Which AI model has the highest ARC-AGI-1 score?

+
6 systems share it at 98.5%: Claude Fable 5, Claude Opus 5.5, Gemini 3.8 Flash, Gemini 4 Argon, GPT-6 Astra and GPT-6.1 Sol. The next score is Gemini 3.1 Pro Preview at 98%. Each row in the table links to the page that published its number.
Q · 02

What is the cheapest model near the top of ARC-AGI-1?

+
DeepSeek V4.1 Flash scores 94.5%, 4.0 points behind the leader, at $0.15 per 1M tokens on our effective blend ($0.3 input, $1.2 output). Claude Fable 5 costs $6.82 on the same blend.
Q · 03

Are these ARC-AGI-1 scores independently verified?

+
99 of 101 published scores come from ARC Prize Foundation; the other 2 are vendor reports. Each row carries a badge that says which, and a link to the source.
Q · 04

What does ARC-AGI-1 measure?

+
ARC-AGI-1 is the original Abstraction and Reasoning Corpus: grid puzzles where a model must infer a hidden rule from a few examples and apply it to a new input. A few input and output grid pairs, usually about three, demonstrate a transformation; the model has to produce the output grid for a new test input.
Q · 05

How is the ARC-AGI-1 score calculated?

+
The share of evaluation tasks solved. ARC Prize publishes the cost per task beside every verified score. The test set has 400 public evaluation tasks; 100-task semi-private and private sets.
Q · 06

How current is this ARC-AGI-1 leaderboard?

+
The newest score was published on November 5, 2026. Every row shows the date its source published it, and we last read all of them on November 8, 2026. Prices come from our model price index, where each rate carries the date it was read on the vendor's page.

One weekly digest. Zero noise.

One weekly digest · No spam, ever · Unsubscribe in one click