ARC-AGI-2 benchmark
63 AI models ranked on ARC-AGI-2. Each row shows the model’s score next to its cheapest API price per 1M tokens where it has one, so you can weigh quality against cost.
Top model today
GPT-5.6 Sol (OpenAI) leads at 93%, from $2.00 per 1M input OpenRouter routed quote.
Data via ARC Prize →Models ranked
63
on this benchmark
Top score
93%
current leader
Type
Single eval
benchmark
Updated
Oct 14, 2026
benchmark scores
Cost vs ARC-AGI-2
61 priced models = reasoning model (extended thinking)
Each labelled name links to that model’s page. Curated to the cost-efficiency frontier, top scorers and cheapest so labels stay readable; the full field is in the leaderboard table below (and via “select all”). Reasoning models (extended “thinking” before answering) carry a ring around the dot.
Models ranked on ARC-AGI-2
highest score first · 63 modelsMaker
| # | ||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | GPT-5.6 Sol |  OpenAI | 93%max$1.44 / task | $2.00 Routed quote | $10.00 Routed quote | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 2 | GPT-5.5 | OpenAI | 85%xhigh$1.87 / task | $5.00 via OpenAI | $30.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 3 | GPT-5.5 Pro | OpenAI | 85%high$11 / task | $30.00 via Aihubmix | $180.00 via Aihubmix | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 4 | GPT-5.6 Terra | OpenAI | 84%max$1.09 / task | $2.00 via OpenAI | $12.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 5 | GPT-5.4 Pro | OpenAI | 83%xhigh$16 / task | $30.00 Routed quote | $180.00 Routed quote | |||||||||||||||||||||||||
| 6 | Gemini 3.1 Pro Preview |  Google | 77% | $2.00 via Google | $12.00 via Google | |||||||||||||||||||||||||
| 7 | Claude Opus 4.7 |  Anthropic | 76%max$7.43 / task | $5.00 via Anthropic | $25.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 8 | GPT-5.4 | OpenAI | 74%xhigh$1.52 / task | $2.50 via Databricks | $15.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 9 | Claude Opus 4.8 | Anthropic | 72%high$2.74 / task | $5.00 via Anthropic | $25.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 10 | Gemini 3.5 Flash | Google | 72%high$0.85 / task | $1.50 via Google | $9.00 via Google | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 11 | Claude Opus 4.6 | Anthropic | 69%high$3.47 / task | $5.00 via Anthropic | $25.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 12 | GPT 5.4 High | OpenAI | 68% | $2.50 via Aihubmix | $15.00 via Aihubmix | |||||||||||||||||||||||||
| 13 | Grok 4 20 Beta 0309 Reasoning |  xAI | 65% | $1.25 via xAI | $2.50 via xAI | |||||||||||||||||||||||||
| 14 | Claude Sonnet 4.6 | Anthropic | 60%high$2.70 / task | $3.00 via Databricks | $15.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 15 | GPT-5.6 Luna | OpenAI | 60%max$0.67 / task | $0.20 via OpenAI | $1.20 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 16 | GPT-5.2 Pro | OpenAI | 54%high$16 / task | $21.00 Routed quote | $168.00 Routed quote | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 17 | GPT 5.2 (2025-12-11) | OpenAI | 53%xhigh$1.90 / task | $1.75 via OpenAI | $14.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 18 | Gemini 3 Deep Think | Google | 45% | n/a | n/a | |||||||||||||||||||||||||
| 19 | Qwen3.8 27B |  Alibaba | 42% | $0.04 Routed quote | $1.49 via Cerebras | |||||||||||||||||||||||||
| 20 | Claude Opus 4.5 (2025-11-01) | Anthropic | 38%reasoning: true · thinking_budget: 64000$2.40 / task | $5.00 via Anthropic | $25.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 21 | Gemini 3 Flash Preview | Google | 34%high$0.231 / task | $0.50 via Google | $3.00 via Google | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 22 | Gemini 3 Pro Preview | Google | 31% | $2.00 via Google | $12.00 via Google | |||||||||||||||||||||||||
| 23 | GPT 5.4 Low | OpenAI | 29% | $2.50 via Aihubmix | $15.00 via Aihubmix | |||||||||||||||||||||||||
| 24 | GLM 5.2 |  Zhipu | 23% | $0.05 Routed quote | $1.98 via Scx AI | |||||||||||||||||||||||||
| 25 | GPT-5.4 Mini | OpenAI | 19%xhigh$0.75 / task | $0.75 via Databricks | $4.50 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 26 | GPT 5.1 (2025-11-13) | OpenAI | 18%high$1.17 / task | $1.25 via OpenAI | $10.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 27 | Grok 4 0709 | xAI | 16% | $1.25 via xAI | $2.50 via xAI | |||||||||||||||||||||||||
| 28 | Claude Sonnet 4.5 (2025-09-29) | Anthropic | 14%reasoning: true · thinking_budget: 32000$0.759 / task | $3.00 via Anthropic | $15.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 29 | Kimi K2.5 |  Moonshot | 12% | $0.45 via DeepInfra | $2.25 via DeepInfra | |||||||||||||||||||||||||
| 30 | GPT 5 (2025-08-07) | OpenAI | 10%high$0.73 / task | $1.25 via OpenAI | $10.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 31 | Claude Opus 4 (2025-05-14) | Anthropic | 9%reasoning: true · thinking_budget: 16000$1.93 / task | $15.00 via Anthropic | $75.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 32 | O3 (2025-04-16) | OpenAI | 7%high$0.834 / task | $2.00 via OpenAI | $8.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 33 | O4 Mini (2025-04-16) | OpenAI | 6%high$0.856 / task | $1.10 via OpenAI | $4.40 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 34 | Claude Sonnet 4 (2025-05-14) | Anthropic | 6%reasoning: true · thinking_budget: 16000$0.486 / task | $3.00 via Anthropic | $15.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 35 | GPT-5.4 Nano | OpenAI | 6%xhigh$0.16 / task | $0.20 via Databricks | $1.25 via Databricks | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 36 | Grok 4 Fast Reasoning | xAI | 5% | $0.20 via Azure | $0.50 via Azure | |||||||||||||||||||||||||
| 37 | MiniMax M2.5 |  MiniMax | 5% | $0.27 Routed quote | $1.08 Routed quote | |||||||||||||||||||||||||
| 38 | GLM 5 | Zhipu | 5% | $0.60 via DeepInfra | $1.92 Routed quote | |||||||||||||||||||||||||
| 39 | Gemini 2.5 Pro | Google | 5%reasoning: true · thinking_budget: 32000$0.757 / task | $1.25 via Databricks | $10.00 via Databricks | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 40 | o3 Pro | OpenAI | 5%high$7.55 / task | $20.00 Routed quote | $80.00 Routed quote | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 41 | GPT 5 Mini (2025-08-07) | OpenAI | 4%high$0.198 / task | $0.25 via OpenAI | $2.00 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 42 | DeepSeek V3.2 |  DeepSeek | 4% | $0.259 Routed quote | $0.38 via DeepInfra | |||||||||||||||||||||||||
| 43 | Claude Haiku 4.5 (2025-10-01) | Anthropic | 4%reasoning: true · thinking_budget: 32000$0.377 / task | $1.00 via Anthropic | $5.00 via Anthropic | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 44 | O3 Mini (2025-01-31) | OpenAI | 3%high$0.547 / task | $1.10 via OpenAI | $4.40 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 45 | GPT 5 Nano (2025-08-07) | OpenAI | 3%high$0.0295 / task | $0.05 via OpenAI | $0.40 via OpenAI | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 46 | Gemini 2.0 Flash | Google | 1% | $0.10 via Google | $0.40 via Google | |||||||||||||||||||||||||
| 47 | Qwen3 235B A22b Instruct 2507 | Alibaba | 1% | $0.09 via DeepInfra | $0.10 via Wandb | |||||||||||||||||||||||||
| 48 | R1 0528 | DeepSeek | 1% | $0.20 via Lambda AI | $0.25 via Hyperbolic | |||||||||||||||||||||||||
| 49 | Claude 3.7 Sonnet | Anthropic | 1%reasoning: true · thinking_budget: 8000$0.36 / task | $3.00 via Databricks | $15.00 via Replicate | |||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
| 50 | O1 Mini (2024-09-12) | OpenAI | 1% | $1.10 via Azure | $4.40 via Azure | |||||||||||||||||||||||||
1–50 of 64
Page 1 of 2
On-demand listed rates, OpenRouter routed quotes and Vercel reseller quotes. Enterprise commitments can be materially lower. Cost per task uses best-effort token usage. An OpenRouter routed quote is the price of whichever upstream provider OpenRouter prefers today, so it moves day to day.
Frequently asked questions
What is the ARC-AGI-2 benchmark?
ARC-AGI-2 is the second-generation Abstraction and Reasoning Corpus from ARC Prize, a test of fluid intelligence rather than recalled knowledge. Each task shows a few input-output grid examples and the model must infer the underlying transformation and apply it to a new grid. The tasks resist memorisation, so frontier models still score far below the human panel. Scored as the percentage of tasks solved on the semi-private evaluation set, which ARC Prize reports alongside each system's cost per task.
Which AI model scores highest on ARC-AGI-2?
GPT-5.6 Sol (OpenAI) leads with 93%, from $2.00 per 1M input tokens as an OpenRouter routed quote, while OpenAI lists it at $4.00 per 1M input.
How many models are ranked on ARC-AGI-2?
63 models carry an ARC-AGI-2 score. Scores come from ARC Prize; where a model has a listed API price it is drawn from LiteLLM and OpenRouter, and models with no current price are listed without one.
Other benchmarks
compare the same models on a different evalIntelligence Indexcomposite indexCoding Indexcomposite indexAgentic Indexcomposite indexGPQA DiamondleaderboardHumanity's Last ExamleaderboardMMLU-ProleaderboardMMMU-ProleaderboardLiveCodeBenchleaderboardSciCodeleaderboardTerminalBench HardleaderboardTerminalBench v2.1leaderboardτ²-benchleaderboardτ-bench BankingleaderboardAPEX AgentsleaderboardIT-Bench SREleaderboardGDPvalleaderboardCritPtleaderboardOmniscienceleaderboardIFBenchleaderboardLong-Context Reasoningcomposite indexSWE-bench VerifiedleaderboardSWE-bench ProleaderboardAIME 2026leaderboardAA-Briefcaseleaderboard
Benchmark scores via ARC Prize · scores updated Oct 14, 2026 · Listed prices from LiteLLM, OpenRouter and Vercel · catalogue refreshed Oct 14, 2026 · 57 prices older, back to Jul 19, 2026
Every weekday
AI moves fast. Here's your debrief.
News, analysis, tools, and more.
For people who build with AI