ARC-AGI

🏆 Leaderboard

As of August 20, 2026, Claude Fable 5 is #1 for ARC-AGI at 98.5%. Ranked by the ARC-AGI-1 Verified score ARC-AGI leaderboard for ARC-AGI 2 and ARC-AGI 3 abstract reasoning scores, plus API pricing.

Updated August 20, 2026282 models33 providers


Claude Fable 5 Anthropic · Proprietary	98.5%	89.2%	—	$60.00
Gemini 3.1 Pro Google · Proprietary	98%	77.1%	0.4%	$17.50
Claude Opus 5 Anthropic · Proprietary	97.5%	90.4%	30.2%	$30.00
GPT-5.5 Pro OpenAI · Proprietary	96.5%	84.6%	—	$540.00
Kimi K3OSS Moonshot AI · Open Source	94.5%	60.4%	—	$18.00
Claude Opus 4.7 Anthropic · Proprietary	93.5%	75.8%	—	$30.00
Gemini 3.6 Flash Google · Proprietary	91.2%	60.4%	—	$4.50
DeepSeek-V4-Flash-0731OSS DeepSeek · Open Source	89%	61.4%	—	$1.76
Claude Opus 4.8 Anthropic · Proprietary	88%	72.1%	1.5%	$30.00
Grok 4.6 xAI · Proprietary	87%	67.1%	2.1%	$8.00
Claude Sonnet 4.6 Anthropic · Proprietary	86%	58.3%	—	$18.00
Claude Opus 4.6 Anthropic · Proprietary	86%	68.8%	—	$30.00
Grok 4.5 xAI · Proprietary	85.7%	52.6%	0.3%	$8.00
Inkling-SmallOSS Thinking Machines · Open Source · via Thinking Machines Lab	84%	40.1%	—	$1.50
GPT-5.2 Pro OpenAI · Proprietary	81.2%	54.2%	—	$189.00
Claude Opus 4.5 Anthropic · Proprietary	80%	37.6%	—	$30.00
GLM-5.2OSS Z AI · Open Source	77%	22.8%	—	$5.80
GPT-5.4 OpenAI · Proprietary	76.7%	73.3%	0.2%	$17.50
GPT-5.5 OpenAI · Proprietary	76.2%	85%	0.4%	$35.00
Gemini 3 Pro Google · Proprietary	75%	31.1%	—	$14.00
GPT-5.6 Sol OpenAI · Proprietary	74.5%	92.5%	7.8%	$35.00
GPT-5.1 High OpenAI · Proprietary	72.8%	—	—	$11.25
GPT-5.3 Codex OpenAI · Proprietary	66.7%	—	—	$15.75
Grok-4 xAI · Proprietary	66.7%	15.9%	—	$18.00
GPT-5 High OpenAI · Proprietary	65.7%	9.9%	—	$11.25
Kimi K2.5OSS Moonshot AI · Open Source	65.3%	11.8%	—	$3.68
MiniMax M2.5OSS MiniMax · Open Source	63.7%	4.9%	—	$1.50
GPT-5.4 mini OpenAI · Proprietary	63.7%	18.9%	—	$5.25
GPT-5.6 Terra OpenAI · Proprietary	60.2%	83.9%	0.8%	$14.00
GPT-5.1 Medium OpenAI · Proprietary	57.7%	—	—	$11.25
DeepSeek-V3.2OSS DeepSeek · Open Source · via OpenRouter	57%	4.0%	—	$0.57
GPT-5 Medium OpenAI · Proprietary	56.2%	7.5%	—	$11.25
GPT-5 mini OpenAI · Proprietary	54.3%	4.4%	—	$2.25
Gemini 3.5 Flash-Lite Google · Proprietary	53.5%	10.3%	—	$2.80
GPT-5.4 nano OpenAI · Proprietary	51.5%	5.7%	—	$1.45
Gemini 3.5 Flash Google · Proprietary	48.8%	72.1%	—	$10.50
Grok-4 Fast Reasoning xAI · Proprietary	48.5%	5.3%	—	$0.70
Grok 4 Fast xAI · Proprietary	48.5%	5.3%	—	$0.70
GLM-5OSS Z AI · Open Source	44.7%	4.9%	—	$4.20
o3-pro OpenAI · Proprietary	44.3%	4.9%	—	$100.00
GPT-5 OpenAI · Proprietary	44%	9.9%	—	$11.25
o3 OpenAI · Proprietary	41.5%	6.5%	—	$10.00
GPT-5.6 Luna OpenAI · Proprietary	34.2%	59.5%	0.2%	$1.40
Gemini 2.5 Flash Google · Proprietary	33.3%	1.7%	—	$2.80
Gemini 2.5 Pro Preview 06-05 Google · Proprietary	31.3%	0%	—	$11.25
o1 OpenAI · Proprietary	30.7%	—	—	$75.00
Claude 3.7 Sonnet Anthropic · Proprietary	28.6%	0.9%	—	$18.00
Claude Sonnet 4.5 Anthropic · Proprietary	25.5%	13.6%	—	$18.00
Claude Sonnet 4 Anthropic · Proprietary	23.8%	5.9%	—	$18.00
o1-pro OpenAI · Proprietary	23.3%	—	—	$750.00

Showing 1–50 of 282 models

Next step

You found the model. Now ship the product.

Auth, billing, and the API layer are already decided. Fork 8 finished AI apps, or have us build the first version with you.

Get the stack — $249 Have us build it with you

The index

All Large Language Models

282 models across 33 providers. Search or jump to a lab — every model page stays linked here.

Cursor

2 models

Composer 2 Composer 2 Fast

xAI

15 models

Grok 4.6 Grok 4.5 Grok 4.3 Grok-4.1 Grok-4.1 Fast Non-Reasoning Grok-4.1 Fast Reasoning Grok-4.1 Thinking Grok 4 Fast Grok Code Fast 1 Grok-4 Fast Non-Reasoning Grok-4 Fast Reasoning Grok-4 Grok-3 Grok-3 Mini Grok-2

DeepSeek

16 models

DeepSeek-V4-Pro-0813 DeepSeek-V4-Flash-0731 DeepSeek-V4-Flash-0423 DeepSeek-V4-Flash-Max DeepSeek-V4-Pro-Max DeepSeek-V3.2 DeepSeek-V3.2 (Non-thinking)DeepSeek-V3.2-Exp DeepSeek-R1-0528 DeepSeek-V3 0324 DeepSeek R1 Distill Llama 70B DeepSeek R1 Distill Qwen 32B DeepSeek-R1 DeepSeek-V3.1 DeepSeek-V3 DeepSeek-V2.5

MiniMax

6 models

MiniMax M3 MiniMax M2.7 MiniMax M2.5 MiniMax M2.1 MiniMax M2 MiniMax M1 80K

AI21 Labs

2 models

Jamba 1.5 Large Jamba 1.5 Mini

Amazon

5 models

Nova 2 Lite Nova 2 Sonic Nova Lite Nova Micro Nova Pro

Baidu

1 models

ERNIE 4.5

ByteDance

2 models

Seed 2.1 Turbo Seed 2.0 Pro

Cohere

3 models

North Mini Code 1.0 Command A+Command R+

IBM

1 models

Granite 3.3 8B Instruct

Inception

1 models

Mercury 2

inclusionAI

1 models

Ling-3.0-flash

LG AI Research

1 models

K-EXAONE-236B-A23B

Meituan

4 models

LongCat-Flash-Lite LongCat-Flash-Thinking-2601 LongCat-Flash-Thinking LongCat-Flash-Chat

Microsoft

5 models

MAI-Code-1.1-Flash Phi 4 Mini Phi-4-multimodal-instruct Phi 4 Phi-3.5-mini-instruct

Moonshot AI

9 models

Kimi K3 Kimi K2.7 Code Kimi K2.7 Code HighSpeed Kimi K2.6 Kimi K2.5 Kimi K2 0905 Kimi K2-Instruct-0905 Kimi K2-Thinking-0905 Kimi K2 Instruct

Nous Research

1 models

Hermes 3 70B

NVIDIA

5 models

Nemotron 3.5 Lightning (30B A3B)Nemotron 3 Ultra (550B A55B)Nemotron 3 Super (120B A12B)Nemotron 3 Nano (30B A3B)Nemotron Nano 9B v2

Poolside

2 models

Laguna S 2.1 Laguna XS 2.1

Sakana AI

1 models

Sakana Namazu

StepFun

1 models

Step-3.5-Flash

Tencent

1 models

Hy3

Thinking Machines

1 models

Inkling-Small

Unisound

1 models

Upstage

1 models

Solar Pro 4

Xiaomi

3 models

MiMo-V2.5-Pro MiMo-V2.5 MiMo-V2-Flash

Which model leads ARC-AGI right now?

As of August 20, 2026, Claude Fable 5 by Anthropic is #1 for ARC-AGI at 98.5%. Ranked by the ARC-AGI-1 Verified score This board also tracks ARC-AGI-1 Verified, ARC-AGI 2, ARC-AGI-3, ARC-AGI-2 Verified. Next on the same board: Gemini 3.1 Pro and Claude Opus 5. Related leaders: GPT-5.6 Sol on ARC-AGI 2 at 92.5%; Claude Opus 5 on ARC-AGI-3 at 30.2%. This arc-agi leaderboard ranks models by ARC-AGI-1 Verified. Scores come from public evals. Prices are the live API rates in the table above.

Sources: ARC-AGI / ARC Prize (https://arcprize.org/); OpenAI API pricing (https://developers.openai.com/api/docs/pricing); Anthropic Claude API pricing (https://platform.claude.com/docs/en/about-claude/pricing); Gemini API pricing (https://ai.google.dev/gemini-api/docs/pricing)

Top 8 for arc-agi. Ranked by the ARC-AGI-1 Verified score Input and output are dollars per million tokens.
Rank	Model	ARC-AGI-1 Verified	Input /M	Output /M
1	Claude Fable 5	98.5%	$10.00	$50.00
2	Gemini 3.1 Pro	98%	$2.50	$15.00
3	Claude Opus 5	97.5%	$5.00	$25.00
4	GPT-5.5 Pro	96.5%	$60.00	$480.00
5	Kimi K3	94.5%	$3.00	$15.00
6	Claude Opus 4.7	93.5%	$5.00	$25.00
7	Gemini 3.6 Flash	91.2%	$0.75	$3.75
8	DeepSeek-V4-Flash-0731	89%	$0.44	$1.32

Related LLM leaderboards

LLM Leaderboard Coding Math Agentic Reasoning Writing Science Finance Long Context Vision Open Source

Individual benchmark pages

Rank one eval at a time. All LLM benchmarks.

HLE SWE Bench Humanity's Last Exam (No Tools)Humanity's Last Exam (With Tools)Terminal-Bench 2.0 Terminal-Bench Terminal-Bench 2.1 Terminal-Bench 2 Terminal-Bench 3.0 ARC-AGI-1 Verified

ARC-AGI FAQ

Who ranks #1 on the ARC-AGI leaderboard?

As of August 20, 2026, Claude Fable 5 by Anthropic ranks #1 on ARC-AGI-1 Verified at 98.5%. API pricing is $10.00/M input and $50.00/M output.

What are the top models on ARC-AGI-1 Verified?

The current ARC-AGI-1 Verified ranking as of August 20, 2026 is 1. Claude Fable 5 at 98.5%; 2. Gemini 3.1 Pro at 98%; 3. Claude Opus 5 at 97.5%.

Which arc-agi model is the cheapest?

Llama 4 Scout is the cheapest scored model on this arc-agi leaderboard at $0.08/M input and $0.30/M output ($0.38 blended). Claude Fable 5 still leads ARC-AGI-1 Verified at 98.5%.

Should I always pick the #1 ARC-AGI-1 Verified model?

Not automatically. Claude Fable 5 leads ARC-AGI-1 Verified, but a cheaper scored model can be the better production choice if the quality gap is small. Use the table to weigh ARC-AGI-1 Verified against input/output price, context window, and related evals.

How often is the ARC-AGI leaderboard updated?

Scores and API prices on this page are refreshed from published evals and provider rates. The snapshot is labeled August 20, 2026. Treat it as a current index, not a one-off blog post.

What is ARC-AGI?

ARC-AGI is François Chollet's Abstraction and Reasoning Corpus. It scores fluid intelligence on novel visual puzzles instead of memorized exams.

What is the difference between ARC-AGI 2 and ARC-AGI 3?

ARC-AGI 2 is the current hard public set. ARC-AGI 3 is the newer generation. We keep both when a lab published a score.

Why are there fewer ARC-AGI scores than GPQA scores?

ARC-AGI is expensive to run and still unpublished for many models. This page ranks the models that have a public number, then shows price.

You found the model. Now ship the product.

All Large Language Models

OpenAI

Anthropic

Google

Cursor

xAI

DeepSeek

MiniMax

Qwen

AI21 Labs

Amazon

Baidu

ByteDance

Cohere

IBM

Inception

inclusionAI

LG AI Research

Meituan

Meta

Microsoft

Mistral

Moonshot AI

Nous Research

NVIDIA

Poolside

Sakana AI

StepFun

Tencent

Thinking Machines

Unisound

Upstage

Xiaomi

Z AI

Which model leads ARC-AGI right now?

Related LLM leaderboards

Individual benchmark pages

ARC-AGI FAQ

Who ranks #1 on the ARC-AGI leaderboard?

What are the top models on ARC-AGI-1 Verified?

Which arc-agi model is the cheapest?

Should I always pick the #1 ARC-AGI-1 Verified model?

How often is the ARC-AGI leaderboard updated?

What is ARC-AGI?

What is the difference between ARC-AGI 2 and ARC-AGI 3?

Why are there fewer ARC-AGI scores than GPQA scores?