New: Claude Fable 5.2 eval report is here.Read the report ↗
Compare models

Sources

Original benchmarks, independent evaluations and provider reports behind Rank’s data.

Updated: Data as of:

Independent evaluations

Rank includes results and measurements published by Artificial Analysis ↗, alongside benchmark publishers and model providers. Their evaluation references are identified throughout the lists below.

Agent Index

Published evaluations used by the current composite index. View the index ↗

HLE-Diamond

  • Artificial Analysis evaluations
    27 reported results · 17 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • HLE-Diamond results
      /models/claude-fable-5-1
    • HLE-Diamond results
      /models/claude-opus-5-5
    • HLE-Diamond results
      /models/claude-sonnet-5-5
    • HLE-Diamond results
      /models/deepseek-v4-1-flash-1017
    • HLE-Diamond results
      /models/gemini-3-8-flash
    • HLE-Diamond results
      /models/gemini-4-argon
    • HLE-Diamond results
      /models/glm-5-3
    • HLE-Diamond results
      /models/glm-5-3-flash
    • HLE-Diamond results
      /models/gpt-6-1-sol
    • HLE-Diamond results
      /models/gpt-6-astra
    • HLE-Diamond results
      /models/grok-4-7-high
    • HLE-Diamond results
      /models/kimi-k3
    • HLE-Diamond results
      /models/kimi-k3-1-pro
    • HLE-Diamond results
      /models/minimax-m3-1-flash
    • HLE-Diamond results
      /models/muse-spark-1-3
    • hle-diamond-aa expanded public leaderboard
      /evaluations/hle-diamond

GDP.pdf

  • Artificial Analysis evaluations
    27 reported results · 8 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • GDP.pdf results
      /evaluations/gdp-pdf
    • GDP.pdf results
      /models/deepseek-v4-1-flash-1017
    • GDP.pdf results
      /models/glm-5-3
    • GDP.pdf results
      /models/glm-5-3-flash
    • GDP.pdf results
      /models/kimi-k3
    • GDP.pdf results
      /models/kimi-k3-1-pro
    • GDP.pdf results
      /models/minimax-m3-1-flash

AA-Briefcase

  • Artificial Analysis evaluations
    27 reported results · 8 source references
    View source references
    • AA-Briefcase results
      /evaluations/aa-briefcase
    • AA-Briefcase results
      /models/deepseek-v4-1-flash-1017
    • AA-Briefcase results
      /models/glm-5-3
    • AA-Briefcase results
      /models/glm-5-3-flash
    • AA-Briefcase results
      /models/kimi-k3
    • AA-Briefcase results
      /models/kimi-k3-1-pro
    • AA-Briefcase results
      /models/minimax-m3-1-flash
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2

Terminal-Bench 4.0

  • Artificial Analysis evaluations
    27 reported results · 17 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • Terminal-Bench results
      /models/claude-fable-5-1-xhigh
    • Terminal-Bench results
      /models/claude-opus-5-5
    • Terminal-Bench results
      /models/claude-sonnet-5-5
    • Terminal-Bench results
      /models/deepseek-v4-1-flash-1017
    • Terminal-Bench results
      /models/gemini-3-8-flash-medium
    • Terminal-Bench results
      /models/gemini-4-argon
    • Terminal-Bench results
      /models/glm-5-3
    • Terminal-Bench results
      /models/glm-5-3-flash
    • Terminal-Bench results
      /models/gpt-6-1-sol
    • Terminal-Bench results
      /models/gpt-6-astra-xhigh
    • Terminal-Bench results
      /models/grok-4-7
    • Terminal-Bench results
      /models/kimi-k3-1-pro
    • Terminal-Bench results
      /models/kimi-k3-low
    • Terminal-Bench results
      /models/minimax-m3-1-flash
    • Terminal-Bench results
      /models/muse-spark-1-3
    • terminal-bench-4-0-aa expanded public leaderboard
      /evaluations/terminalbench-4-0

Benchmarks and reports

AA-Briefcase

  • Artificial Analysis evaluations
    67 reported results · 22 source references
    View source references
    • AA-Briefcase results
      /evaluations/aa-briefcase
    • AA-Briefcase results
      /models/deepseek-r1-0120
    • AA-Briefcase results
      /models/deepseek-v4-1-flash
    • AA-Briefcase results
      /models/deepseek-v4-1-flash-1017
    • AA-Briefcase results
      /models/deepseek-v4-flash-0420-high
    • AA-Briefcase results
      /models/deepseek-v4-pro
    • AA-Briefcase results
      /models/deepseek-v4-pro-0424
    • AA-Briefcase results
      /models/glm-5-2
    • AA-Briefcase results
      /models/glm-5-3
    • AA-Briefcase results
      /models/glm-5-3-flash
    • AA-Briefcase results
      /models/gpt-oss-120b
    • AA-Briefcase results
      /models/gpt-oss-20b
    • AA-Briefcase results
      /models/kimi-k2-6
    • AA-Briefcase results
      /models/kimi-k3
    • AA-Briefcase results
      /models/kimi-k3-1-pro
    • AA-Briefcase results
      /models/minimax-m3
    • AA-Briefcase results
      /models/minimax-m3-1-flash
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
  • Anthropic · 4 reported results
  • Thinking Machines Lab · 1 reported result

AutomationBench-AA

  • Artificial Analysis evaluations
    69 reported results · 65 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • AutomationBench-AA public evaluation chart
      /models/claude-4-5-sonnet-thinking
    • AutomationBench-AA public evaluation chart
      /models/claude-fable-5
    • AutomationBench-AA public evaluation chart
      /models/claude-fable-5-1
    • AutomationBench-AA public evaluation chart
      /models/claude-haiku-5-5-xhigh
    • AutomationBench-AA public evaluation chart
      /models/claude-opus-4-8
    • AutomationBench-AA public evaluation chart
      /models/claude-opus-5
    • AutomationBench-AA public evaluation chart
      /models/claude-opus-5-5
    • AutomationBench-AA public evaluation chart
      /models/claude-sonnet-4-6-adaptive
    • AutomationBench-AA public evaluation chart
      /models/claude-sonnet-5
    • AutomationBench-AA public evaluation chart
      /models/gemini-3-1-pro-preview
    • AutomationBench-AA public evaluation chart
      /models/gemini-3-8-flash-medium
    • AutomationBench-AA public evaluation chart
      /models/glm-5-1
    • AutomationBench-AA public evaluation chart
      /models/gpt-5-4-mini
    • AutomationBench-AA public evaluation chart
      /models/gpt-5-5
    • AutomationBench-AA public evaluation chart
      /models/gpt-5-6-luna
    • AutomationBench-AA public evaluation chart
      /models/gpt-5-6-sol
    • AutomationBench-AA public evaluation chart
      /models/gpt-5-6-terra
    • AutomationBench-AA public evaluation chart
      /models/gpt-6-1-sol-xhigh
    • AutomationBench-AA public evaluation chart
      /models/gpt-6-1-terra
    • AutomationBench-AA public evaluation chart
      /models/gpt-6-astra
    • AutomationBench-AA public evaluation chart
      /models/gpt-6-luna
    • AutomationBench-AA public evaluation chart
      /models/grok-4-7
    • AutomationBench-AA public evaluation chart
      /models/inkling
    • AutomationBench-AA public evaluation chart
      /models/mimo-v2-6-pro
    • AutomationBench-AA public evaluation chart
      /models/minimax-m2-7
    • AutomationBench-AA public evaluation chart
      /models/muse-spark-1-1
    • AutomationBench-AA public evaluation chart
      /models/muse-spark-1-2
    • AutomationBench-AA public evaluation chart
      /models/muse-spark-1-3
    • AutomationBench-AA public evaluation chart
      /models/muse-spark-open
    • AutomationBench-AA public evaluation chart
      /models/muse-wonder
    • AutomationBench-AA public evaluation chart
      /models/nemotron-3-5-super-vl
    • AutomationBench-AA public evaluation chart
      /models/qwen3-14b-instruct-reasoning
    • AutomationBench-AA public evaluation chart
      /models/qwen3-5-397b-a17b
    • AutomationBench-AA public evaluation chart
      /models/qwen3-5-9b
    • AutomationBench-AA public evaluation chart
      /models/qwen3-8-27b
    • AutomationBench-AA public evaluation chart
      /models/step-5
    • AutomationBench-AA results
      /evaluations/automationbench-aa
    • automationbench-aa-1-0-6 published result
      /models/deepseek-r1-0120
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-1-flash
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-1-flash-1017
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-1-flash-1017-high
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-1-flash-1017-non-reasoning
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-1-flash-non-reasoning
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-flash-0420
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-flash-0420-high
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-pro
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-pro-0424
    • automationbench-aa-1-0-6 published result
      /models/deepseek-v4-pro-non-reasoning
    • automationbench-aa-1-0-6 published result
      /models/glm-5-2
    • automationbench-aa-1-0-6 published result
      /models/glm-5-3
    • automationbench-aa-1-0-6 published result
      /models/glm-5-3-flash
    • automationbench-aa-1-0-6 published result
      /models/glm-5-3-low
    • automationbench-aa-1-0-6 published result
      /models/gpt-oss-120b
    • automationbench-aa-1-0-6 published result
      /models/gpt-oss-20b
    • automationbench-aa-1-0-6 published result
      /models/kimi-k2-6
    • automationbench-aa-1-0-6 published result
      /models/kimi-k3
    • automationbench-aa-1-0-6 published result
      /models/kimi-k3-1-pro
    • automationbench-aa-1-0-6 published result
      /models/kimi-k3-low
    • automationbench-aa-1-0-6 published result
      /models/minimax-m3
    • automationbench-aa-1-0-6 published result
      /models/minimax-m3-1-flash

BrowseComp

FrontierCode

  • Artificial Analysis evaluations
    5 reported results · 5 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
  • Anthropic · 1 reported result
  • Anthropic · 4 reported results

GDP.pdf

  • Artificial Analysis evaluations
    65 reported results · 22 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • GDP.pdf results
      /evaluations/gdp-pdf
    • GDP.pdf results
      /models/deepseek-r1-0120
    • GDP.pdf results
      /models/deepseek-v4-1-flash
    • GDP.pdf results
      /models/deepseek-v4-1-flash-1017
    • GDP.pdf results
      /models/deepseek-v4-flash-0420
    • GDP.pdf results
      /models/deepseek-v4-pro
    • GDP.pdf results
      /models/deepseek-v4-pro-0424
    • GDP.pdf results
      /models/glm-5-2
    • GDP.pdf results
      /models/glm-5-3
    • GDP.pdf results
      /models/glm-5-3-flash
    • GDP.pdf results
      /models/gpt-oss-120b
    • GDP.pdf results
      /models/gpt-oss-20b
    • GDP.pdf results
      /models/kimi-k2-6
    • GDP.pdf results
      /models/kimi-k3
    • GDP.pdf results
      /models/kimi-k3-1-pro
    • GDP.pdf results
      /models/minimax-m3
    • GDP.pdf results
      /models/minimax-m3-1-flash
  • Thinking Machines Lab · 1 reported result
  • surgehq.ai

GDPval-AA

  • Artificial Analysis evaluations
    94 reported results · 95 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • GDPval-AA v2.1 results
      /models/claude-fable-5
    • GDPval-AA v2.1 results
      /models/claude-fable-5-1
    • GDPval-AA v2.1 results
      /models/claude-fable-5-1-high
    • GDPval-AA v2.1 results
      /models/claude-fable-5-1-low
    • GDPval-AA v2.1 results
      /models/claude-fable-5-1-medium
    • GDPval-AA v2.1 results
      /models/claude-fable-5-1-xhigh
    • GDPval-AA v2.1 results
      /models/claude-opus-4-7
    • GDPval-AA v2.1 results
      /models/claude-opus-4-8
    • GDPval-AA v2.1 results
      /models/claude-opus-5
    • GDPval-AA v2.1 results
      /models/claude-opus-5-5
    • GDPval-AA v2.1 results
      /models/claude-opus-5-5-high
    • GDPval-AA v2.1 results
      /models/claude-opus-5-5-low
    • GDPval-AA v2.1 results
      /models/claude-opus-5-5-medium
    • GDPval-AA v2.1 results
      /models/claude-opus-5-5-xhigh
    • GDPval-AA v2.1 results
      /models/claude-opus-5-high
    • GDPval-AA v2.1 results
      /models/claude-opus-5-low
    • GDPval-AA v2.1 results
      /models/claude-opus-5-medium
    • GDPval-AA v2.1 results
      /models/claude-opus-5-xhigh
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-5
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-5-high
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-5-low
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-5-medium
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-5-xhigh
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-high
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-low
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-medium
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-non-reasoning
    • GDPval-AA v2.1 results
      /models/claude-sonnet-5-xhigh
    • GDPval-AA v2.1 results
      /models/deepseek-v4-1-flash
    • GDPval-AA v2.1 results
      /models/deepseek-v4-1-flash-1017
    • GDPval-AA v2.1 results
      /models/deepseek-v4-1-flash-non-reasoning
    • GDPval-AA v2.1 results
      /models/deepseek-v4-pro
    • GDPval-AA v2.1 results
      /models/deepseek-v4-pro-non-reasoning
    • GDPval-AA v2.1 results
      /models/gemini-3-1-pro-preview
    • GDPval-AA v2.1 results
      /models/gemini-3-8-flash
    • GDPval-AA v2.1 results
      /models/gemini-3-8-flash-low
    • GDPval-AA v2.1 results
      /models/gemini-3-8-flash-medium
    • GDPval-AA v2.1 results
      /models/gemini-4-argon
    • GDPval-AA v2.1 results
      /models/gemini-4-carbon
    • GDPval-AA v2.1 results
      /models/glm-5-3
    • GDPval-AA v2.1 results
      /models/glm-5-3-low
    • GDPval-AA v2.1 results
      /models/gpt-5-5
    • GDPval-AA v2.1 results
      /models/gpt-5-5-high
    • GDPval-AA v2.1 results
      /models/gpt-5-5-low
    • GDPval-AA v2.1 results
      /models/gpt-5-5-medium
    • GDPval-AA v2.1 results
      /models/gpt-5-5-non-reasoning
    • GDPval-AA v2.1 results
      /models/gpt-5-6-sol
    • GDPval-AA v2.1 results
      /models/gpt-5-6-sol-high
    • GDPval-AA v2.1 results
      /models/gpt-5-6-sol-low
    • GDPval-AA v2.1 results
      /models/gpt-5-6-sol-medium
    • GDPval-AA v2.1 results
      /models/gpt-5-6-sol-non-reasoning
    • GDPval-AA v2.1 results
      /models/gpt-5-6-sol-xhigh
    • GDPval-AA v2.1 results
      /models/gpt-6-1-sol
    • GDPval-AA v2.1 results
      /models/gpt-6-1-sol-high
    • GDPval-AA v2.1 results
      /models/gpt-6-1-sol-low
    • GDPval-AA v2.1 results
      /models/gpt-6-1-sol-medium
    • GDPval-AA v2.1 results
      /models/gpt-6-1-sol-xhigh
    • GDPval-AA v2.1 results
      /models/gpt-6-astra
    • GDPval-AA v2.1 results
      /models/gpt-6-astra-high
    • GDPval-AA v2.1 results
      /models/gpt-6-astra-low
    • GDPval-AA v2.1 results
      /models/gpt-6-astra-medium
    • GDPval-AA v2.1 results
      /models/gpt-6-astra-xhigh
    • GDPval-AA v2.1 results
      /models/grok-4-7
    • GDPval-AA v2.1 results
      /models/grok-4-7-high
    • GDPval-AA v2.1 results
      /models/grok-4-7-low
    • GDPval-AA v2.1 results
      /models/kimi-k3
    • GDPval-AA v2.1 results
      /models/kimi-k3-1-pro
    • GDPval-AA v2.1 results
      /models/kimi-k3-low
    • GDPval-AA v2.1 results
      /models/minimax-m3-1-flash
    • GDPval-AA v2.1 results
      /models/muse-spark
    • GDPval-AA v2.1 results
      /models/muse-spark-1-1
    • GDPval-AA v2.1 results
      /models/muse-spark-1-2
    • GDPval-AA v2.1 results
      /models/muse-spark-1-3
    • GDPval-AA v2.1 results
      /models/muse-spark-1-3-xhigh
    • gdpval-aa-v2-1 published result
      /models/deepseek-r1-0120
    • gdpval-aa-v2-1 published result
      /models/deepseek-v4-flash-0420
    • gdpval-aa-v2-1 published result
      /models/deepseek-v4-flash-0420-high
    • gdpval-aa-v2-1 published result
      /models/deepseek-v4-pro-0424
    • gdpval-aa-v2-1 published result
      /models/deepseek-v4-pro-0424-high
    • gdpval-aa-v2-1 published result
      /models/glm-5-2
    • gdpval-aa-v2-1 published result
      /models/glm-5-2-non-reasoning
    • gdpval-aa-v2-1 published result
      /models/glm-5-3-flash
    • gdpval-aa-v2-1 published result
      /models/gpt-oss-120b
    • gdpval-aa-v2-1 published result
      /models/gpt-oss-120b-low
    • gdpval-aa-v2-1 published result
      /models/gpt-oss-20b
    • gdpval-aa-v2-1 published result
      /models/kimi-k2-5
    • gdpval-aa-v2-1 published result
      /models/kimi-k2-6
    • gdpval-aa-v2-1 published result
      /models/minimax-m3
    • GDPval-AA official benchmark
      /evaluations/gdpval-aa
  • Thinking Machines Lab · 1 reported result

GPQA Diamond

  • Artificial Analysis evaluations
    67 reported results · 67 source references
    View source references
    • GPQA Diamond results
      /models/claude-fable-5
    • GPQA Diamond results
      /models/claude-fable-5-1
    • GPQA Diamond results
      /models/claude-fable-5-1-high
    • GPQA Diamond results
      /models/claude-fable-5-1-low
    • GPQA Diamond results
      /models/claude-fable-5-1-medium
    • GPQA Diamond results
      /models/claude-fable-5-1-xhigh
    • GPQA Diamond results
      /models/claude-opus-4-6
    • GPQA Diamond results
      /models/claude-opus-4-6-adaptive
    • GPQA Diamond results
      /models/claude-opus-4-7
    • GPQA Diamond results
      /models/claude-opus-4-7-non-reasoning
    • GPQA Diamond results
      /models/claude-opus-4-8
    • GPQA Diamond results
      /models/claude-opus-5
    • GPQA Diamond results
      /models/claude-opus-5-high
    • GPQA Diamond results
      /models/claude-opus-5-low
    • GPQA Diamond results
      /models/claude-opus-5-medium
    • GPQA Diamond results
      /models/claude-opus-5-xhigh
    • GPQA Diamond results
      /models/claude-sonnet-5
    • GPQA Diamond results
      /models/claude-sonnet-5-non-reasoning
    • GPQA Diamond results
      /models/deepseek-v4-pro
    • GPQA Diamond results
      /models/gemini-3-1-pro-preview
    • GPQA Diamond results
      /models/gemini-3-8-flash
    • GPQA Diamond results
      /models/gemini-3-8-flash-low
    • GPQA Diamond results
      /models/gemini-3-8-flash-medium
    • GPQA Diamond results
      /models/glm-5-3
    • GPQA Diamond results
      /models/gpt-5-5
    • GPQA Diamond results
      /models/gpt-5-5-high
    • GPQA Diamond results
      /models/gpt-5-5-low
    • GPQA Diamond results
      /models/gpt-5-5-medium
    • GPQA Diamond results
      /models/gpt-5-5-non-reasoning
    • GPQA Diamond results
      /models/gpt-5-6-sol
    • GPQA Diamond results
      /models/gpt-5-6-sol-high
    • GPQA Diamond results
      /models/gpt-5-6-sol-low
    • GPQA Diamond results
      /models/gpt-5-6-sol-medium
    • GPQA Diamond results
      /models/gpt-5-6-sol-non-reasoning
    • GPQA Diamond results
      /models/gpt-5-6-sol-xhigh
    • GPQA Diamond results
      /models/gpt-6-astra
    • GPQA Diamond results
      /models/gpt-6-astra-high
    • GPQA Diamond results
      /models/gpt-6-astra-low
    • GPQA Diamond results
      /models/gpt-6-astra-medium
    • GPQA Diamond results
      /models/gpt-6-astra-xhigh
    • GPQA Diamond results
      /models/kimi-k3
    • GPQA Diamond results
      /models/kimi-k3-low
    • GPQA Diamond results
      /models/muse-spark
    • GPQA Diamond results
      /models/muse-spark-1-1
    • GPQA Diamond results
      /models/muse-spark-1-2
    • GPQA Diamond results
      /models/muse-spark-1-3
    • GPQA Diamond results
      /models/muse-spark-1-3-xhigh
    • gpqa-diamond-aa published result
      /models/deepseek-r1-0120
    • gpqa-diamond-aa published result
      /models/deepseek-v4-flash-0420
    • gpqa-diamond-aa published result
      /models/deepseek-v4-flash-0420-high
    • gpqa-diamond-aa published result
      /models/deepseek-v4-flash-0420-non-reasoning
    • gpqa-diamond-aa published result
      /models/deepseek-v4-pro-0424
    • gpqa-diamond-aa published result
      /models/deepseek-v4-pro-0424-high
    • gpqa-diamond-aa published result
      /models/deepseek-v4-pro-0424-non-reasoning
    • gpqa-diamond-aa published result
      /models/glm-5-2
    • gpqa-diamond-aa published result
      /models/glm-5-2-non-reasoning
    • gpqa-diamond-aa published result
      /models/glm-5-3-flash
    • gpqa-diamond-aa published result
      /models/gpt-oss-120b
    • gpqa-diamond-aa published result
      /models/gpt-oss-120b-low
    • gpqa-diamond-aa published result
      /models/gpt-oss-20b
    • gpqa-diamond-aa published result
      /models/gpt-oss-20b-low
    • gpqa-diamond-aa published result
      /models/kimi-k2-5
    • gpqa-diamond-aa published result
      /models/kimi-k2-5-non-reasoning
    • gpqa-diamond-aa published result
      /models/kimi-k2-6
    • gpqa-diamond-aa published result
      /models/kimi-k2-6-non-reasoning
    • gpqa-diamond-aa published result
      /models/kimi-k2-thinking
    • gpqa-diamond-aa published result
      /models/minimax-m3
  • Ricursive Intelligence · 1 reported result
  • Thinking Machines Lab · 1 reported result
  • Thinking Machines Lab · 1 reported result
  • github.com

Harvey LAB-AA

  • Artificial Analysis evaluations
    3 reported results · 2 source references
    View source references
    • Harvey LAB-AA v1.1 launch results
      /articles/harvey-lab-aa-v1-1
    • Harvey LAB-AA official benchmark
      /evaluations/harvey-lab-aa
  • Artificial Analysis via Anthropic system card · 5 reported results

HLE-Diamond

  • Artificial Analysis evaluations
    40 reported results · 28 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • HLE-Diamond results
      /models/claude-fable-5-1
    • HLE-Diamond results
      /models/claude-opus-5-5
    • HLE-Diamond results
      /models/claude-sonnet-5-5
    • HLE-Diamond results
      /models/deepseek-v4-1-flash
    • HLE-Diamond results
      /models/deepseek-v4-1-flash-1017
    • HLE-Diamond results
      /models/deepseek-v4-pro
    • HLE-Diamond results
      /models/gemini-3-1-pro-preview
    • HLE-Diamond results
      /models/gemini-3-8-flash
    • HLE-Diamond results
      /models/gemini-4-argon
    • HLE-Diamond results
      /models/gemini-4-carbon
    • HLE-Diamond results
      /models/glm-5-3
    • HLE-Diamond results
      /models/glm-5-3-flash
    • HLE-Diamond results
      /models/gpt-6-1-sol
    • HLE-Diamond results
      /models/gpt-6-astra
    • HLE-Diamond results
      /models/gpt-oss-120b
    • HLE-Diamond results
      /models/gpt-oss-20b
    • HLE-Diamond results
      /models/grok-4-7-high
    • HLE-Diamond results
      /models/kimi-k3
    • HLE-Diamond results
      /models/kimi-k3-1-pro
    • HLE-Diamond results
      /models/minimax-m3
    • HLE-Diamond results
      /models/minimax-m3-1-flash
    • HLE-Diamond results
      /models/muse-spark-1-3
    • hle-diamond-aa expanded public leaderboard
      /evaluations/hle-diamond
  • Thinking Machines Lab · 1 reported result
  • lastexam.ai

Humanity’s Last Exam

  • Artificial Analysis evaluations
    94 reported results · 93 source references
    View source references
    • Artificial Analysis original HLE score
      /models/gpt-6-1-sol
    • Artificial Analysis original HLE score
      /models/gpt-6-1-sol-high
    • Artificial Analysis original HLE score
      /models/gpt-6-1-sol-low
    • Artificial Analysis original HLE score
      /models/gpt-6-1-sol-medium
    • Artificial Analysis original HLE score
      /models/gpt-6-1-sol-xhigh
    • Artificial Analysis original HLE score
      /models/gpt-6-astra
    • Artificial Analysis original HLE score
      /models/gpt-6-astra-high
    • Artificial Analysis original HLE score
      /models/gpt-6-astra-low
    • Artificial Analysis original HLE score
      /models/gpt-6-astra-medium
    • Artificial Analysis original HLE score
      /models/gpt-6-astra-xhigh
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • Humanity's Last Exam Benchmark
      /evaluations/humanitys-last-exam
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-4-5-sonnet-thinking
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-fable-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-fable-5-1
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-haiku-5-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-opus-4-5-thinking
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-opus-4-6-adaptive
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-opus-4-7
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-opus-4-8
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-opus-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-opus-5-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-sonnet-4-6-adaptive
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-sonnet-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/claude-sonnet-5-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/deepseek-v3-2-reasoning
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/deepseek-v4-1-flash
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/deepseek-v4-1-flash-1017
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/deepseek-v4-pro
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gemini-3-1-pro-preview
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gemini-3-8-flash
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gemini-3-flash-reasoning
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gemini-3-pro
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/glm-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/glm-5-1
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/glm-5-3
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-4-1
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-4-1-mini
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-4-turbo
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-2
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-4
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-4-mini
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-5
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-6-luna
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-6-sol
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-5-6-terra
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-6-1-terra
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/gpt-6-luna
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/inkling
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/kimi-k3
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/kimi-k3-1-pro
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/mimo-v2-6-pro
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/minimax-m2-7
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/minimax-m3-1-flash
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/muse-spark
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/muse-spark-1-1
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/muse-spark-1-2
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/muse-spark-1-3
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/muse-spark-open
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/muse-wonder
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/nemotron-3-5-super-vl
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/qwen3-14b-instruct-reasoning
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/qwen3-5-35b-a3b
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/qwen3-5-397b-a17b
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/qwen3-5-9b
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/qwen3-6-plus
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/qwen3-8-27b
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/step-3-7-flash
    • Humanity's Last Exam: independent original-HLE evaluation
      /models/step-5
    • hle-text published result
      /models/deepseek-r1-0120
    • hle-text published result
      /models/deepseek-v4-flash-0420
    • hle-text published result
      /models/deepseek-v4-flash-0420-high
    • hle-text published result
      /models/deepseek-v4-flash-0420-non-reasoning
    • hle-text published result
      /models/deepseek-v4-pro-0424
    • hle-text published result
      /models/deepseek-v4-pro-0424-high
    • hle-text published result
      /models/deepseek-v4-pro-0424-non-reasoning
    • hle-text published result
      /models/glm-5-2
    • hle-text published result
      /models/glm-5-2-non-reasoning
    • hle-text published result
      /models/glm-5-3-flash
    • hle-text published result
      /models/gpt-oss-120b
    • hle-text published result
      /models/gpt-oss-120b-low
    • hle-text published result
      /models/gpt-oss-20b
    • hle-text published result
      /models/gpt-oss-20b-low
    • hle-text published result
      /models/kimi-k2-5
    • hle-text published result
      /models/kimi-k2-5-non-reasoning
    • hle-text published result
      /models/kimi-k2-6
    • hle-text published result
      /models/kimi-k2-6-non-reasoning
    • hle-text published result
      /models/kimi-k2-thinking
    • hle-text published result
      /models/minimax-m3
  • Anthropic · 2 reported results
  • Anthropic · 2 reported results
  • Anthropic · 1 reported result
  • Anthropic · 4 reported results
  • Anthropic · 6 reported results
  • Anthropic · 3 reported results
  • Anthropic · 3 reported results
  • Anthropic · 2 reported results
  • Anthropic · 4 reported results
  • DeepSeek · 7 reported results
  • DeepSeek · 2 reported results
  • Meta · 1 reported result
  • Meta · 4 reported results
  • Moonshot AI · 4 reported results
  • Moonshot AI · 3 reported results
  • Moonshot AI · 10 reported results
  • Moonshot AI · 6 reported results
  • Moonshot AI · 1 reported result
  • OpenAI · 1 reported result
  • OpenAI · 12 reported results
  • Thinking Machines Lab · 1 reported result
  • Thinking Machines Lab · 2 reported results
  • Z.ai · 2 reported results
  • Z.ai · 3 reported results
  • Z.ai · 1 reported result
  • Z.ai · 1 reported result
  • xAI · 1 reported result
  • lastexam.ai

IFBench

  • Artificial Analysis evaluations
    17 reported results · 11 source references
    View source references
    • IFBench results
      /evaluations/ifbench
    • IFBench results
      /models/deepseek-r1-0120
    • IFBench results
      /models/deepseek-v4-flash-0420
    • IFBench results
      /models/deepseek-v4-pro-0424
    • IFBench results
      /models/glm-5-2
    • IFBench results
      /models/gpt-oss-120b
    • IFBench results
      /models/gpt-oss-20b
    • IFBench results
      /models/kimi-k2-5
    • IFBench results
      /models/kimi-k2-6
    • IFBench results
      /models/kimi-k2-thinking
    • IFBench results
      /models/minimax-m3
  • Ricursive Intelligence · 1 reported result
  • Thinking Machines Lab · 1 reported result
  • github.com

SWE-bench Verified

Terminal-Bench

  • Artificial Analysis evaluations
    79 reported results · 47 source references
    View source references
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-high
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-low
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-medium
    • Artificial Analysis published evaluation
      /models/claude-fable-5-2-xhigh
    • Terminal-Bench results
      /models/claude-fable-5
    • Terminal-Bench results
      /models/claude-fable-5-1-xhigh
    • Terminal-Bench results
      /models/claude-opus-4-8
    • Terminal-Bench results
      /models/claude-opus-5
    • Terminal-Bench results
      /models/claude-opus-5-5
    • Terminal-Bench results
      /models/claude-sonnet-5
    • Terminal-Bench results
      /models/claude-sonnet-5-5
    • Terminal-Bench results
      /models/deepseek-r1-0120
    • Terminal-Bench results
      /models/deepseek-v4-1-flash
    • Terminal-Bench results
      /models/deepseek-v4-1-flash-1017
    • Terminal-Bench results
      /models/deepseek-v4-flash-0420-high
    • Terminal-Bench results
      /models/deepseek-v4-pro
    • Terminal-Bench results
      /models/deepseek-v4-pro-0424
    • Terminal-Bench results
      /models/gemini-3-1-pro-preview
    • Terminal-Bench results
      /models/gemini-3-8-flash-medium
    • Terminal-Bench results
      /models/gemini-4-argon
    • Terminal-Bench results
      /models/gemini-4-carbon
    • Terminal-Bench results
      /models/glm-5-2
    • Terminal-Bench results
      /models/glm-5-3
    • Terminal-Bench results
      /models/glm-5-3-flash
    • Terminal-Bench results
      /models/gpt-5-5
    • Terminal-Bench results
      /models/gpt-5-6-sol
    • Terminal-Bench results
      /models/gpt-6-1-sol
    • Terminal-Bench results
      /models/gpt-6-astra-xhigh
    • Terminal-Bench results
      /models/gpt-oss-120b
    • Terminal-Bench results
      /models/gpt-oss-20b
    • Terminal-Bench results
      /models/grok-4-7
    • Terminal-Bench results
      /models/kimi-k2-6
    • Terminal-Bench results
      /models/kimi-k3-1-pro
    • Terminal-Bench results
      /models/kimi-k3-low
    • Terminal-Bench results
      /models/minimax-m3
    • Terminal-Bench results
      /models/minimax-m3-1-flash
    • Terminal-Bench results
      /models/muse-spark-1-1
    • Terminal-Bench results
      /models/muse-spark-1-2
    • Terminal-Bench results
      /models/muse-spark-1-3
    • terminal-bench-2-1-aa published result
      /models/deepseek-v4-flash-0420
    • terminal-bench-2-1-aa published result
      /models/deepseek-v4-pro-0424-high
    • terminal-bench-2-1-aa published result
      /models/glm-5-2-non-reasoning
    • terminal-bench-2-1-aa published result
      /models/gpt-oss-120b-low
    • terminal-bench-2-1-aa published result
      /models/kimi-k2-5
    • terminal-bench-2-1-aa published result
      /models/kimi-k3
    • terminal-bench-4-0-aa expanded public leaderboard
      /evaluations/terminalbench-4-0
  • Anthropic · 2 reported results
  • Cognition · 2 reported results
  • DeepSeek · 3 reported results
  • MiniMax · 1 reported result
  • Moonshot AI · 1 reported result
  • Ricursive Intelligence · 1 reported result
  • Terminal-Bench · 19 reported results
  • Thinking Machines Lab · 3 reported results
  • Thinking Machines Lab · 1 reported result
  • Z.ai · 1 reported result
  • Z.ai · 1 reported result

Terminal-Bench-Science

Model specifications and provider measurements