Find the right AI for you

Search by name or tell us what you want to do
API servicesChoose how to access modelsAI modelsCapabilities, specifications, and pricingAI toolsFind useful tools by taskMCP directoryConnect tools, data, and workflowsAgent skillsFind reusable skills for AI agentsBenchmarksCompare model performanceOnline toolsConvert accounts, format JSON, estimate costs and test APIsModel timelineExplore model releasesUse casesStart with a practical task
EnterSearch · ↑↓Choose a resultEscClose
Submit中文
Project benchmark

ARC-AGI-3 · standard framework

ARC Prize · 70 configurations with published scores. Review each metric and its test conditions separately.

Report identity & conditions

Data as of2026-11-03· Report published—. See individual records for run dates and test conditions.

Leaderboard updated

  • Run configurations remain separate, including different reasoning settings.
  • Individual test dates are not given. Update dates and model release dates from the report remain separate and do not substitute for actual test dates.
  • Costs follow the report’s run scope and dollar amounts; they are not model API unit prices.
  • ARC-AGI-3 · Semi-private evaluation set. Benchmark versions are compared separately.
  • Learning efficiency for interactive tasks. A score of 100% matches human efficiency at completing the games; this is not static-question accuracy.
  • Total evaluation cost in USD. It is not converted into cost per task and is not a current API quote.
  • The standard harness uses a minimal, provider-neutral interface and allows the model to keep notes.
Read the original report ↗

Compare metrics within the same report. Each chart explains its direction and scope.
See the details for full values and run conditions.

Skill acquisition efficiency

Learning efficiency for interactive tasks. A score of 100% matches human efficiency at completing the games; this is not static-question accuracy.: showing 8 of 70 records with values. See details for configurations and dates.

Chart data · 8 entries
Name%
GPT-6 Astra (Max)62.713
GPT-6 Astra (XHigh)59.343
GPT-6 Astra (High)54.819
GPT-6.1 Sol (Max)52.729
Gemini 4 Argon (Low)49.794
Gemini 4 Argon (Medium)46.266
GPT-6.1 Sol (XHigh)39.929
GPT-6 Astra (Medium)38.588

Total run cost

Total evaluation cost in USD. It is not converted into cost per task and is not a current API quote.: showing 8 of 70 records with values. See details for configurations and dates.

Chart data · 8 entries
NameUSD
GPT-6 Luna (Low)228.833
GPT-6 Luna (None)229.771
GPT-6 Luna (Medium)241.28
GPT-6 Luna (High)255.979
GPT-6 Luna (Max)262.238
GPT-6 Luna (XHigh)262.684
Claude Haiku 5.5 (Max)388.106
DeepSeek V4.1 Flash (None)734.725

Full results & model versions

70 / 70configurations have published scores in this scope. Scored configurations are listed; missing values remain “Not provided.” Different run configurations of the same model are listed separately.

Claude Sonnet 5.5 (Medium)

Model reported by the benchmark

Standard framework

As reportedOrganization: Anthropic

anthropic-claude-sonnet-5-5 · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.3961083341784243%
Total run cost
4839.68249 USD

Anthropic Opus 4.6 (Max)

Model reported by the benchmark

Standard framework

As reportedOrganization: Anthropic

anthropic-opus-4-6-max-effort · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.51%
Total run cost
8866.2 USD

Muse Spark Open (XHigh)

Model reported by the benchmark

Standard framework

As reportedOrganization: Meta AI

meta-muse-spark-open · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.13621973920554634%
Total run cost
3908.42107525 USD

GPT-6 Astra (Max)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-astra · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
62.71280210060628%
Total run cost
26097.501720000007 USD

Claude Sonnet 5.5 (High)

Model reported by the benchmark

Standard framework

As reportedOrganization: Anthropic

anthropic-claude-sonnet-5-5 · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
1.203076515291301%
Total run cost
5487.227968 USD

GPT-6 Sol (High)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-sol · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.899438991115236%
Total run cost
5072.447082 USD

DeepSeek V4.1 Flash (None)

Model reported by the benchmark

Standard framework

As reportedOrganization: DeepSeek

deepseek-v4-1-flash · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.3431694493873915%
Total run cost
734.7252066 USD

GPT-6 Luna (Medium)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-luna · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.1941528463935471%
Total run cost
241.2797313 USD

GPT-6 Sol (None)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-sol · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.2837742350643006%
Total run cost
4655.549962 USD

GPT-6 Astra (None)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-astra · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
35.177880317424%
Total run cost
49791.11317000001 USD

GPT-6 Astra (High)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-astra · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
54.81905163612231%
Total run cost
40704.722830000006 USD

Grok 4.5 (Medium)

Model reported by the benchmark

Standard framework

As reportedOrganization: xAI

xai-grok-4-5 · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.32%
Total run cost
8458.3 USD

Claude Sonnet 5.5 (Low)

Model reported by the benchmark

Standard framework

As reportedOrganization: Anthropic

anthropic-claude-sonnet-5-5 · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.4047513036164488%
Total run cost
4756.680228 USD

GPT-5.6 Luna (Medium)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-5-6-luna · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.17%
Total run cost
2379.13 USD

GPT-6 Sol (XHigh)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-sol · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
1.7976269602783517%
Total run cost
5105.623836 USD

GPT-5.6 Luna (Low)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-5-6-luna · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.17%
Total run cost
2307.49 USD

GPT-6 Astra (XHigh)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-astra · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
59.34295409146055%
Total run cost
37317.38768 USD

GPT-5.4 (High)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-5-4-2026-03-05-thinking · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.21%
Total run cost
5187.41 USD

GPT-6 Sol (Medium)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-6-sol · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.418612306617149%
Total run cost
4863.795116 USD

GPT-5.5 (High)

Model reported by the benchmark

Standard framework

As reportedOrganization: OpenAI

openai-gpt-5-5-2026-04-23-thinking · CoT

Release time reported by the benchmark:

Skill acquisition efficiency
0.43%
Total run cost
10000 USD