Anthropic
Best on the Capabilities Index Claude Fable 5.2 173.4
How does Anthropic stack up?
Every model on Epoch AI's Capabilities Index by release date, with Anthropic's in color.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)
Anthropic's best models
Anthropic Nov 2026
Claude Fable 5.2
Closed weights Flagship
173.4
#1 of 256
Epoch Capabilities Index
Self-Reported Index 175.3 #1 of 404
- AutomationBench 51.5% #2 /42
- GPQA Diamond 93.9% #18 /229
- HealthBench Professional 67.7% #3 /26
Anthropic Sep 2026
Claude Opus 5.5
Closed weights Flagship
168.5
#4 of 256
Epoch Capabilities Index
Self-Reported Index 170.0 #5 of 404
- AutomationBench 42.5% #6 /42
- Agents' Last Exam 38.2% #1 /31
- Terminal-Bench 4.0 64.8% #1 /40
Anthropic Sep 2026
Claude Fable 5.1
Closed weights Flagship
167.0
#7 of 256
Epoch Capabilities Index
Self-Reported Index 167.9 #8 of 404
- Humanity's Last Exam 46.5% #2 /47
- AutomationBench 22.4% #15 /42
- Terminal-Bench 4.0 57.9% #6 /40
Anthropic Sep 2026
Claude Sonnet 5.5
Closed weights Mid-size
166.6
#8 of 256
Epoch Capabilities Index
Self-Reported Index 168.0 #7 of 404
- Humanity's Last Exam 46.1% #4 /47
- AutomationBench 44.8% #5 /42
- Terminal-Bench 4.0 61.8% #2 /40
Anthropic Jul 2026
Claude Opus 5
Closed weights Flagship
164.7
#10 of 256
Epoch Capabilities Index
Self-Reported Index 166.8 #10 of 404
- DeepSWE 73.7% #3 /28
- AutomationBench 26.9% #12 /42
- Agents' Last Exam 32.2% #3 /31
Anthropic Jun 2026
Claude Fable 5
Closed weights Flagship
164.1
#11 of 256
Epoch Capabilities Index
Self-Reported Index 162.1 #20 of 404
- DeepSWE 69.9% #5 /28
- AutomationBench 17.1% #19 /42
- Agents' Last Exam 25.7% #15 /31
What benchmarks does Anthropic mention most?
Here are the 28 benchmarks Anthropic mentions most in its launch posts.
HumanEval MMLU GSM8K GPQA Diamond MMMU SWE-bench Verified OSWorld-Verified tau-bench MMMLU Terminal-Bench AIME 2025 tau2-bench Vals Finance Agent ARC-AGI-2 Terminal-Bench 2.0 MCP Atlas Humanity's Last Exam BrowseComp GDPval-AA SWE-Bench Pro SWE-bench Multilingual SWE-bench Multimodal AutomationBench HealthBench Professional FrontierCode GDPval-AA v2.1 Chartography AA-Briefcase v1.1
Source: Labs’ launch posts, read by Benchmark Almanac
BBQ (Bias Benchmark for QA) GPQA Diamond SWE-bench Verified Malicious computer use (Anthropic) Child safety evaluations (Anthropic) Virology Capabilities Test (VCT) Long-form virology tasks LAB-Bench FigQA Quadruped reinforcement learning task (Anthropic) LLM training task (Anthropic) Cybench Single-turn harmful request evaluations (Anthropic) Single-turn benign request evaluations (Anthropic) Malicious use of Claude Code (Anthropic) Anthropic automated behavioral audit Multi-turn testing (Anthropic) SHADE-Arena DNA Synthesis Screening Evasion Impossible tasks (Anthropic) Kernels task (Anthropic) Anthropic political even-handedness eval Prompt injection robustness against adaptive attackers, coding (Anthropic) Prompt injection robustness against adaptive attackers, computer use (Anthropic) SWE-bench Multilingual Humanity's Last Exam Prompt injection robustness against adaptive attackers, browser use (Anthropic) SWE-Bench Pro Suicide and self-harm evaluations (Anthropic)
Source: Labs’ system cards and technical reports, read by Benchmark Almanac
BBQ (Bias Benchmark for QA) GPQA Diamond MMMU SWE-bench Verified OSWorld-Verified Malicious computer use (Anthropic) Child safety evaluations (Anthropic) Virology Capabilities Test (VCT) MMMLU Long-form virology tasks LAB-Bench FigQA Single-turn harmful request evaluations (Anthropic) Single-turn benign request evaluations (Anthropic) Malicious use of Claude Code (Anthropic) Anthropic automated behavioral audit Multi-turn testing (Anthropic) SHADE-Arena DNA Synthesis Screening Evasion Anthropic political even-handedness eval Humanity's Last Exam SWE-Bench Pro SWE-bench Multilingual Prompt injection robustness against adaptive attackers, coding (Anthropic) Prompt injection robustness against adaptive attackers, computer use (Anthropic) BrowseComp SWE-bench Multimodal AutomationBench HealthBench Professional
Source: Labs’ launch posts, system cards and technical reports, read by Benchmark Almanac
Model cards
Every Anthropic model
- Claude Fable 5.2 Nov 2026
- Claude Haiku 5.5 Oct 2026
- Claude Sonnet 5.5 Sep 2026
- Claude Opus 5.5 Sep 2026
- Claude Fable 5.1 Sep 2026
- Claude Opus 5 Jul 2026
- Claude Sonnet 5 Jun 2026
- Claude Fable 5 Jun 2026
- Claude Opus 4.8 May 2026
- Claude Opus 4.7 Apr 2026
- Claude Mythos Preview Apr 2026
- Claude Sonnet 4.6 Feb 2026
- Claude Opus 4.6 Feb 2026
- Claude Opus 4.5 Nov 2025
- Claude Haiku 4.5 Oct 2025
- Claude Sonnet 4.5 Sep 2025
- Claude Opus 4.1 Aug 2025
- Claude Opus 4 May 2025
- Claude Sonnet 4 May 2025
- Claude 3.7 Sonnet Feb 2025
- Claude 3.5 Haiku Oct 2024
- Claude 3.5 Sonnet (October 2024) Oct 2024
- Claude 3.5 Sonnet Jun 2024
- Claude 3 Haiku Mar 2024
- Claude 3 Opus Feb 2024
- Claude 3 Sonnet Feb 2024
- Claude 2.1 Nov 2023
- Claude Instant Aug 2023
- Claude 2 Jul 2023
- Claude 1.3 Apr 2023