Models / Anthropic

Anthropic / MODEL EVIDENCE

Claude Opus 5.5

127 benchmarks · 292 observations · 102 sources · released 2026-09-22

Scores at a glance

119 benchmarks · large dot: Claude Opus 5.5; small dots: closest models where reported

Artificial Analysis Intelligence Index v5.0 ↗58.41 index

Claude Opus 5.5 (Max, Default Fallback) · Intelligence Index v5.0 ↗

2 more reported results

57.62 index · Claude Opus 5.5 (Max, Default Fallback) · Intelligence Index v4.3.2 · Artificial Analysis ↗

58 · Claude Opus 5.5 (max) · AA Index · Meta report snapshot · Muse Wonder official evaluation report ↗

AA-Briefcase v1.1 ↗1799.65 Elo

Claude Opus 5.5 (Max, Default Fallback) · AA-Briefcase v1.1 · overall Elo ↗

1 more reported result

1803 Elo · Claude Opus 5.5 (max, with fallback) · AA-Briefcase v1.1 · AA-Briefcase v1.1 (Elo-score) · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

AutomationBench-AA ↗69.54%

Claude Opus 5.5 (Max, Default Fallback) · AutomationBench-AA · guardrail-gated objective score ↗

2 more reported results

41.2 · Claude Opus 5.5 (max) · AutomationBench-AA · Muse Wonder official evaluation report ↗

69.5% · Claude Opus 5.5 (max, with fallback) · AutomationBench-AA · Automation Bench-AA · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

GDP.pdf ↗26.20%

Claude Opus 5.5 (Max, Default Fallback) · GDP.pdf · all-pass ↗

4 more reported results

30.6 · Claude Opus 5.5 (max) · GDP.pdf · Muse Wonder official evaluation report ↗

30.6% · Claude Opus 5.5 (max) · GDP.pdf · GDP.pdf score · % · Epoch AI · GDP.pdf ↗

83.9% · Claude Opus 5.5 (max, with fallback) · GDP.pdf · GDP.pdf w/ python · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

30.6% · Claude Opus 5.5 (max, with fallback) · GDP.pdf · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

Terminal-Bench 4.0 ↗59.60%

Claude Opus 5.5 (Max, Default Fallback) · Terminal-Bench 4.0 · AA harness ↗

5 more reported results

65.15 % · Claude Opus 5.5 (Mini-SWE-agent) · Terminal-Bench 4.0 · Score · Vals AI · Terminal-Bench 4.0 ↗

66.4% · Claude Opus 5.5 · Terminal-Bench 4.0 · Gemini 4 Argon model card ↗

66.4% · Opus 5.5 · Terminal-Bench 4.0 · Claude Opus 5.5 launch report ↗

59.6 · Claude Opus 5.5 (max) · Terminal-Bench 4.0-AA · Muse Wonder official evaluation report ↗

66.4% · Claude Opus 5.5 (max, with fallback) · Terminal-Bench 4.0 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

Terminal-Bench-Science 0.1 ↗59.05%

Claude Opus 5.5 (Max, Default Fallback) · Terminal-Bench-Science 0.1 · AA harness ↗

4 more reported results

47.14 % · Claude Opus 5.5 (Mini-SWE-agent) · Terminal-Bench-Science · Score · Vals AI · Terminal-Bench-Science ↗

63.3% · Claude Opus 5.5 · Terminal-Bench Science 0.1 · Gemini 4 Argon model card ↗

58.7% · Opus 5.5 · Terminal-Bench-Science 0.1 · Claude Opus 5.5 launch report ↗

63.3% · Claude Opus 5.5 (max, with fallback) · Terminal-Bench-Science 0.1 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

FrontierCode 2.0 ↗46.44%

Claude Opus 5.5 (Max, Default Fallback) · FrontierCode 2.0 · AA implementation ↗

CritPt v2 · hard subset ↗61.00%

Claude Opus 5.5 (Max, Default Fallback) · CritPt v2 · hard subset · AA implementation ↗

HLE-Diamond ↗67.90%

Claude Opus 5.5 (Max, Default Fallback) · HLE-Diamond · AA implementation ↗

2 more reported results

55.00% · claude-opus-5-5 · Scale Labs · Score · Scale Labs / SEAL · Scale Labs ↗

60.9% · Claude Opus 5.5 (max, with fallback) · HLE-Diamond · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

AA-Omniscience v1.1 ↗48.13 index

Claude Opus 5.5 (Max, Default Fallback) · AA-Omniscience v1.1 · error-penalized index ↗

1 more reported result

46.42 index · Claude Opus 5.5 (Max, Default Fallback) · AA-Omniscience · error-penalized score · Artificial Analysis ↗

AA-Omniscience v1.1 · accuracy ↗65.50%

Claude Opus 5.5 (Max, Default Fallback) · AA-Omniscience v1.1 · accuracy ↗

1 more reported result

66.22% · Claude Opus 5.5 (Max, Default Fallback) · AA-Omniscience · accuracy · Artificial Analysis ↗

AA-Omniscience v1.1 · hallucination rate ↗50.34%

Claude Opus 5.5 (Max, Default Fallback) · AA-Omniscience v1.1 · hallucination rate, lower is better ↗

1 more reported result

58.61% · Claude Opus 5.5 (Max, Default Fallback) · AA-Omniscience · hallucination rate, lower is better · Artificial Analysis ↗

GDPval-AA v2.1 ↗1866.27 Elo

Claude Opus 5.5 (Max, Default Fallback) · GDPval-AA v2.1 · Elo ↗

2 more reported results

1846 · Opus 5.5 · GDPval-AA v2.1 (Elo) · Claude Opus 5.5 launch report ↗

1866 Elo · Claude Opus 5.5 (max, with fallback) · GDPval-AA v2.1 · GDPval-AA v2.1 (Elo-score) · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

AA-3D ↗1196.40 Elo

Claude Opus 5.5 (Max, Default Fallback) · AA-3D · Elo ↗

AA-AnalystAgent ↗56.25%

Claude Opus 5.5 (Max, Default Fallback) · AA-AnalystAgent · strict score ↗

Harvey LAB-AA v1.1 ↗4.17%

Claude Opus 5.5 (Max, Default Fallback) · Harvey LAB-AA v1.1 · all-pass ↗

1 more reported result

4.2 · Claude Opus 5.5 (max) · Harvey LAB-AA · Muse Wonder official evaluation report ↗

FrontierCode 1.1 ↗52.17%

Claude Opus 5.5 (Max, Default Fallback) · FrontierCode 1.1 · AA implementation ↗

2 more reported results

54.4% · Opus 5.5 · FrontierCode 1.1 Main · Claude Opus 5.5 launch report ↗

54.64% · Claude Opus 5.5 (medium) · FrontierCode · Main score · % · Epoch AI · FrontierCode ↗

AA-LCR v1.1 ↗84.67%

Claude Opus 5.5 (Max, Default Fallback) · AA-LCR v1.1 ↗

SciCode ↗66.90%

Claude Opus 5.5 (Max, Default Fallback) · SciCode · AA implementation ↗

5 more reported results

66.9% · Claude Opus 5.5 (max) · SciCode · Score · % · Epoch AI · SciCode ↗

65.05% · Claude Opus 5.5 (xhigh) · SciCode · Score · % · Epoch AI · SciCode ↗

60.42% · Claude Opus 5.5 (high) · SciCode · Score · % · Epoch AI · SciCode ↗

59.26% · Claude Opus 5.5 (medium) · SciCode · Score · % · Epoch AI · SciCode ↗

58.56% · Claude Opus 5.5 (low) · SciCode · Score · % · Epoch AI · SciCode ↗

Humanity’s Last Exam ↗61.35%

Claude Opus 5.5 (Max, Default Fallback) · Humanity’s Last Exam · AA implementation ↗

3 more reported results

67.7% · tools · Opus 5.5 · Humanity’s Last Exam · tools · Claude Opus 5.5 launch report ↗

67.7% · Claude Opus 5.5 (max, with fallback) · Humanity’s Last Exam · HLE-Full w/ tools · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

59.8% · Claude Opus 5.5 (max, with fallback) · Humanity’s Last Exam · HLE-Full · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

ITBench-AA ↗38.23%

Claude Opus 5.5 (Max, Default Fallback) · ITBench-AA · root-cause score ↗

MMMU-Pro ↗87.69%

Claude Opus 5.5 (Max, Default Fallback) · MMMU-Pro · AA implementation ↗

2 more reported results

89.8% · Claude Opus 5.5 (max, with fallback) · MMMU-Pro · MMMU-Pro w/ python · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

87.7% · Claude Opus 5.5 (max, with fallback) · MMMU-Pro · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

MLCR-AA ↗27.78%

Claude Opus 5.5 (Max, Default Fallback) · MLCR-AA · overall ↗

HealthBench ↗59.88%

Claude Opus 5.5 (Max, Default Fallback) · HealthBench · HealthBench Professional · length-adjusted ↗

1 more reported result

65.6 · Claude Opus 5.5 (max) · HealthBench · HealthBench Professional · length-adjusted · Muse Wonder official evaluation report ↗

CritPt v2 ↗77.71%

Claude Opus 5.5 (Max, Default Fallback) · CritPt v2 · code tools ↗

CritPt ↗31.71%

Claude Opus 5.5 (Max, Default Fallback) · CritPt · original ↗

5 more reported results

31.71% · Claude Opus 5.5 (max) · CritPt · Accuracy · % · Epoch AI · CritPt ↗

31.71% · Claude Opus 5.5 (xhigh) · CritPt · Accuracy · % · Epoch AI · CritPt ↗

30.86% · Claude Opus 5.5 (high) · CritPt · Accuracy · % · Epoch AI · CritPt ↗

27.71% · Claude Opus 5.5 (medium) · CritPt · Accuracy · % · Epoch AI · CritPt ↗

17.71% · Claude Opus 5.5 (low) · CritPt · Accuracy · % · Epoch AI · CritPt ↗

AA-PersonalityBench ↗89.85 percentile

Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · expressiveness ↗

7 more reported results

77.15 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · organization · Artificial Analysis ↗

92.38 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · informativeness · Artificial Analysis ↗

89.85 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · transparency · Artificial Analysis ↗

82.23 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · warmth · Artificial Analysis ↗

69.54 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · conviction · Artificial Analysis ↗

94.92 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · imagination · Artificial Analysis ↗

18.77 percentile · Claude Opus 5.5 (Max, Default Fallback) · AA-PersonalityBench · slop · Artificial Analysis ↗

CursorBench 4.0 ↗57.8%

Opus 5.5 · CursorBench 4.0 ↗

5 more reported results

57.8% · Claude Opus 5.5 (max) · CursorBench · Score · % · Epoch AI · CursorBench ↗

56% · Claude Opus 5.5 (xhigh) · CursorBench · Score · % · Epoch AI · CursorBench ↗

56% · Claude Opus 5.5 (high) · CursorBench · Score · % · Epoch AI · CursorBench ↗

52.5% · Claude Opus 5.5 (medium) · CursorBench · Score · % · Epoch AI · CursorBench ↗

43.7% · Claude Opus 5.5 (low) · CursorBench · Score · % · Epoch AI · CursorBench ↗

AutomationBench · Zapier ↗42.5%

Claude Opus 5.5 · AutomationBench ↗

1 more reported result

40.0% · Opus 5.5 · AutomationBench · Claude Opus 5.5 launch report ↗

OSWorld 2.1 ↗81.8% · partial

Opus 5.5 · OSWorld 2.1 · partial ↗

2 more reported results

81.8% partial / — binary · Claude Opus 5.5 (max) · OSWorld 2.1 · partial and binary · Muse Wonder official evaluation report ↗

81.8% · Claude Opus 5.5 (max, with fallback) · OSWorld 2.1 · OSWorld V2.1 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

Chartography ↗66.3%

Claude Opus 5.5 · Chartography · no tools ↗

2 more reported results

89.0% · tools · Opus 5.5 · Chartography · tools · Claude Opus 5.5 launch report ↗

89% · Claude Opus 5.5 (max, with fallback) · Chartography · ChartoGraphy w/ python · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

Vals Index ↗70.18 %

Claude Opus 5.5 · Vals Index · Accuracy ↗

1 more reported result

67.0% · Claude Opus 5.5 · Vals Index · Gemini 4 Argon model card ↗

Vals Finance Agent v2 ↗58.59 %

Claude Opus 5.5 · Finance Agent v2 · Accuracy ↗

3 more reported results

58.6% · Claude Opus 5.5 · Vals Finance Agent v2 · Gemini 4 Argon model card ↗

58.6 · Claude Opus 5.5 (max) · Vals Finance Agent v2 · Muse Wonder official evaluation report ↗

56.7% · Claude Opus 5.5 (max, with fallback) · Vals Finance Agent v2 · Frontier Finance · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

Harvey Legal Agent Benchmark ↗3.75 %

Claude Opus 5.5 · Harvey's Legal Agent Benchmark v1 · Accuracy ↗

1 more reported result

3.8% · Claude Opus 5.5 · Harvey LAB · Vals v1 · Gemini 4 Argon model card ↗

DeepSWE v1.1 ↗74.2%

Claude Opus 5.5 · DeepSWE v1.1 ↗

2 more reported results

74.2 · Claude Opus 5.5 (max) · DeepSWE 1.1 · Muse Wonder official evaluation report ↗

74.2% · Claude Opus 5.5 (max, with fallback) · DeepSWE v1.1 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

FrontierSWE v2 ↗62.3%

Claude Opus 5.5 · FrontierSWE v2 ↗

1 more reported result

62.33% · Claude Opus 5.5 (max) · FrontierSWE · Score · % · Epoch AI · FrontierSWE ↗

Vibe Code Bench ↗30.36 %

Claude Opus 5.5 · Vibe Code Bench · Vibe Code Bench 1–100 ↗

2 more reported results

90.29 % · Claude Opus 5.5 (OpenHands) · Vibe Code Bench · original task set · Vals AI · Vibe Code Bench ↗

90.3% · Claude Opus 5.5 · Vibe Code Bench · version unspecified in launch report · Gemini 4 Argon model card ↗

PostTrainBench ↗49.3%

Claude Opus 5.5 · PostTrainBench ↗

1 more reported result

43.8% · Claude Opus 5.5 (max, with fallback) · PostTrainBench · PostTrain Bench v1.2 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

LABBench 2 ↗73.1%

Claude Opus 5.5 · LABBench 2 ↗

RiemannBench ↗69.6%

Claude Opus 5.5 · RiemannBench ↗

GraphWalks ↗90.6%

Claude Opus 5.5 · GraphWalks BFS ≤128k · F1 ↗

1 more reported result

66.8% · Claude Opus 5.5 · GraphWalks BFS 256k–1M · F1 · Gemini 4 Argon model card ↗

Agents’ Last Exam ↗38.2%

Claude Opus 5.5 · Agents’ Last Exam · pass rate ↗

2 more reported results

42.6% · Claude Opus 5.5 (max, with fallback) · Agents’ Last Exam · Agents' Last Exam (ALL) · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

40.7% · Claude Opus 5.5 (max, with fallback) · Agents’ Last Exam · Agents' Last Exam (CLI) · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

LVBench ↗83.7%

Claude Opus 5.5 · LVBench ↗

CWE-Bench ↗67.0%

Claude Opus 5.5 · CWE-Bench v1 ↗

Personal Superintelligence Index ↗55.2

Claude Opus 5.5 (max) · Personal Superintelligence Index · internal ↗

JobBench ↗75.4

Claude Opus 5.5 (max) · JobBench ↗

MemoryBench ↗81.5

Claude Opus 5.5 (max) · MemoryBench ↗

SWE-Marathon ↗46.3

Claude Opus 5.5 (max) · SWE-Marathon ↗

1 more reported result

46.3% · Claude Opus 5.5 (max, with fallback) · SWE-Marathon · SWE Marathon v1.1 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

ProgramBench ↗18.5 %

Claude Opus 5.5 · ProgramBench · Accuracy ↗

2 more reported results

70.5 · Claude Opus 5.5 (max) · ProgramBench · Muse Wonder official evaluation report ↗

65% · Claude Opus 5.5 (max, with fallback) · ProgramBench · Program Bench · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

STEMBench ↗2.621 / 4

Claude Opus 5.5 · STEMBench120 · mean progress / 4 · closed ↗

ErdősBench ↗3.240 / 4

Claude Opus 5.5 xhigh · ErdősBench · mean progress / 4 · API ↗

GPQA Diamond ↗90.59%

Claude Opus 5.5 (max) · GPQA diamond · Best score (across scorers) · % ↗

1 more reported result

93.9% · Claude Opus 5.5 (max, with fallback) · GPQA Diamond · GPQA-Diamond · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

FrontierMath Tiers 1-3 (v2) ↗91.23%

Claude Opus 5.5 (max) · FrontierMath-Tiers-1-3-v2-Private · Best score (across scorers) · % ↗

FrontierMath Tier 4 (v2) ↗95%

Claude Opus 5.5 (max) · FrontierMath-Tier-4-v2-Private · Best score (across scorers) · % ↗

OTIS Mock AIME 2024-2025 ↗100%

Claude Opus 5.5 (max) · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % ↗

SimpleQA Verified ↗72.2%

Claude Opus 5.5 (max) · SimpleQA Verified · Best score (across scorers) · % ↗

Chess Puzzles ↗67%

Claude Opus 5.5 (max) · Chess Puzzles · Best score (across scorers) · % ↗

Earthborne Rangers (EBR-bench) ↗71.43%

Claude Opus 5.5 (max) · EBR-bench · Best score (across scorers) · % ↗

Mystery Game Puzzles ↗71%

Claude Opus 5.5 (max) · Mystery Game Puzzles · Best score (across scorers) · % ↗

MirrorCode ↗77.4%

Claude Opus 5.5 (max) · MirrorCode · Best score (across scorers) · % ↗

ARC-AGI-1 ↗98.5%

Claude Opus 5.5 (high) · ARC-AGI · Score · % ↗

4 more reported results

97.5% · Claude Opus 5.5 (max) · ARC-AGI · Score · % · Epoch AI · ARC-AGI ↗

97.5% · Claude Opus 5.5 (xhigh) · ARC-AGI · Score · % · Epoch AI · ARC-AGI ↗

97.5% · Claude Opus 5.5 (medium) · ARC-AGI · Score · % · Epoch AI · ARC-AGI ↗

88.5% · Claude Opus 5.5 (low) · ARC-AGI · Score · % · Epoch AI · ARC-AGI ↗

APEX-Agents · original ↗73.5%

Claude Opus 5.5 (max) · APEX-Agents · Pass@1 score · % ↗

2 more reported results

52.5% · Claude Opus 5.5 (medium) · APEX-Agents · Pass@1 score · % · Epoch AI · APEX-Agents ↗

73.5% · Claude Opus 5.5 (max, with fallback) · APEX-Agents · original · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

ARC-AGI-2 ↗93.33%

Claude Opus 5.5 (high) · ARC-AGI-2 · Score · % ↗

4 more reported results

92.5% · Claude Opus 5.5 (xhigh) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

91.67% · Claude Opus 5.5 (max) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

87.5% · Claude Opus 5.5 (medium) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

70.14% · Claude Opus 5.5 (low) · ARC-AGI-2 · Score · % · Epoch AI · ARC-AGI-2 ↗

ProofBench ↗100 %

Claude Opus 5.5 · ProofBench · Accuracy ↗

1 more reported result

100% · Claude Opus 5.5 (max) · ProofBench · Accuracy · % · Epoch AI · ProofBench ↗

LMCA ↗68.23%

Claude Opus 5.5 (max) · LMCA · Score · % ↗

4 more reported results

66.18% · Claude Opus 5.5 (xhigh) · LMCA · Score · % · Epoch AI · LMCA ↗

65.91% · Claude Opus 5.5 (high) · LMCA · Score · % · Epoch AI · LMCA ↗

65.66% · Claude Opus 5.5 (medium) · LMCA · Score · % · Epoch AI · LMCA ↗

64.92% · Claude Opus 5.5 (low) · LMCA · Score · % · Epoch AI · LMCA ↗

DTBench ↗98.93%

Claude Opus 5.5 (max) · DTBench · Accuracy · % ↗

4 more reported results

97.33% · Claude Opus 5.5 (xhigh) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

96.53% · Claude Opus 5.5 (high) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

96.53% · Claude Opus 5.5 (medium) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

95.47% · Claude Opus 5.5 (low) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

FrontierMath: Erdős (Lean 4) ↗2.94%

Claude Opus 5.5 (max) · FrontierMath-Erdos · Best score (across scorers) · % ↗

Furniture Assembly Benchmark ↗83.33%

Claude Opus 5.5 (max) · Furniture Assembly · Best score (across scorers) · % ↗

Text Arena (Coding) ↗1820.31 Elo

Claude Opus 5.5 (max) · WebDev Arena · Elo ↗

WeirdML (v3) ↗49.83%

Claude Opus 5.5 (xhigh) · WeirdML v3 · Max performance · % ↗

SpireClimb ↗3.33%

Claude Opus 5.5 (max) · SpireClimb · Best score (across scorers) · % ↗

Epoch Capabilities Index (ECI) ↗168.83 ECI

Claude Opus 5.5 · General ECI · 26 October 2026 fit ↗

Vals RSI Index ↗0.46 reference score

Claude Opus 5.5 (max) · Compression · reference-anchored score ↗

3 more reported results

0.53 reference score · Claude Opus 5.5 (max) · LM training · reference-anchored score · Vals RSI Index · LM training ↗

0.15 reference score · Claude Opus 5.5 (max) · Harness engineering · reference-anchored score · Vals RSI Index · Harness engineering ↗

0.35 reference score · Claude Opus 5.5 (max) · Post-training · reference-anchored score · Vals RSI Index · Post-training ↗

Vals Finance Index ↗65.59 %

Claude Opus 5.5 · Vals Finance Index · Accuracy ↗

Vals Legal Index ↗72.33 %

Claude Opus 5.5 · Vals Legal Index · Accuracy ↗

Vals Science Index ↗58.65 %

Claude Opus 5.5 · Vals Science Index · Accuracy ↗

Legal Research Bench ↗50.48 %

Claude Opus 5.5 · Legal Research Bench · All-Pass ↗

HLAB-Vals-1.1 ↗9.17 %

Claude Opus 5.5 · HLAB-Vals-1.1 · Task Pass Rate ↗

Vals Finance Agent v2.1 ↗67.85 %

Claude Opus 5.5 · Finance Agent · Accuracy ↗

Excel Modeling Benchmark (EMB & EMB v1.1) ↗75.84 %

Claude Opus 5.5 · Excel Modeling Benchmark · Accuracy ↗

1 more reported result

75.94 % · Claude Opus 5.5 · Excel Modeling Benchmark v1 · Accuracy · Vals AI · Excel Modeling Benchmark v1 ↗

Tax Agent Bench ↗45.57 %

Claude Opus 5.5 · Tax Agent Bench · Accuracy ↗

MedCode ↗49.8 %

Claude Opus 5.5 · MedCode · Accuracy ↗

MedScribe ↗91.43 %

Claude Opus 5.5 · MedScribe · Accuracy ↗

SAFE-Teen ↗27%

Claude Opus 5.5 · Urgent safety · share of applicable ratings marked fail ↗

2 more reported results

44% · Claude Opus 5.5 · Health advice · share of applicable ratings marked fail · Vals AI · SAFE-Teen ↗

25% · Claude Opus 5.5 · Relationship boundaries · share of applicable ratings marked fail · Vals AI · SAFE-Teen ↗

BioMysteryBench ↗79.26 %

Claude Opus 5.5 · BioMysteryBench · Accuracy ↗

1 more reported result

81.9% · Claude Opus 5.5 (max, with fallback) · BioMysteryBench · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

MysteryMechanism ↗49.55 %

Claude Opus 5.5 · MysteryMechanism · Accuracy ↗

Robo-bench ↗14.58 %

Claude Opus 5.5 · Robo-bench · Accuracy ↗

SAGE (Student Assessment with Generative Evaluation) ↗45.83 %

Claude Opus 5.5 · SAGE · Accuracy ↗

Code Migration v1.1 ↗81.01 %

Claude Opus 5.5 · Code Migration · Accuracy ↗

1 more reported result

66.65 % · Claude Opus 5.5 · Code Migration v1 · Accuracy · Vals AI · Code Migration v1 ↗

IOI v1 & IOI v2 ↗95.06 %

Claude Opus 5.5 · IOI · Accuracy ↗

SWE-bench Multimodal ↗70.21 %

Claude Opus 5.5 · SWE-bench Multimodal · Accuracy ↗

CyberBench v1.1 ↗55.36 %

Claude Opus 5.5 · CyberBench · Accuracy ↗

SRE-Bench ↗33.59 %

Claude Opus 5.5 · SRE Bench · Accuracy ↗

CUA-bench ↗14 %

Claude Opus 5.5 · CUA-bench · Accuracy ↗

Time Horizon Index: KSP ↗91.33 %

Claude Opus 5.5 · Time Horizon Index: KSP · Overall ↗

Public Benefits Bench v1.1 ↗70.64 %

Claude Opus 5.5 · Public Benefits Bench · Accuracy ↗

Humanity's Sixth Sense (HSS) ↗44.60%

Claude-Opus-5.5 (xhigh) · Humanity's Sixth Sense · Score ↗

DistressBench ↗75.58%

claude-opus-5.5 · DistressBench · Score ↗

HiL-Bench (Human-in-Loop Benchmark) ↗59.00%

Claude Opus 5.5 · HiL-Bench (Human-in-Loop Benchmark) · Score ↗

PRBench — Legal (Professional Reasoning) ↗47.58%

claude-opus-5-5 · Scale Labs · Score ↗

Remote Labor Index ↗21.25%

Opus 5.5 · Scale Labs · Score ↗

Fortress ↗13.18%

claude-opus-5-5 · Scale Labs · Risk elicitation score ↗

TasteVal ↗2.3 ×

Opus 5.5 (max) · TasteVal · Experimental compute multiplier · October 2026 paper comparison ↗

1 more reported result

1.14 × · Opus 5.5 (max) · TasteVal · Normalized final-performance multiplier · October 2026 paper comparison · TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts ↗

FocusBench ↗45.8%

Claude Opus 5.5 (max) · FocusBench · score · Meta · Muse Wonder launch · maximum effort ↗

BenchCAD ↗73%

Claude Opus 5.5 (max, with fallback) · BenchCAD · BenchCAD w/o tools · Kimi K3.1 Pro · complete release comparison ↗

1 more reported result

96.2% · Claude Opus 5.5 (max, with fallback) · BenchCAD · BenchCAD w/ tools · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

Kimi Code Bench 3.0 ↗58.5%

Claude Opus 5.5 (max, with fallback) · Kimi Code Bench 3.0 · Kimi Code Bench 3.0 Long-horizon (Internal) · Kimi K3.1 Pro · complete release comparison ↗

BrowseComp ↗90.8%

Claude Opus 5.5 (max, with fallback) · BrowseComp · Kimi K3.1 Pro · complete release comparison ↗

CharXiv ↗93.8%

Claude Opus 5.5 (max, with fallback) · CharXiv · CharXiv (RQ) w/ python · Kimi K3.1 Pro · complete release comparison ↗

1 more reported result

92.1% · Claude Opus 5.5 (max, with fallback) · CharXiv · CharXiv (RQ) · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

BabyVision ↗97.7%

Claude Opus 5.5 (max, with fallback) · BabyVision · BabyVision w/ python · Kimi K3.1 Pro · complete release comparison ↗

ZeroBench ↗61%

Claude Opus 5.5 (max, with fallback) · ZeroBench · ZeroBench_main w/ python (pass@5) · Kimi K3.1 Pro · complete release comparison ↗

2 more reported results

35% · Claude Opus 5.5 (max, with fallback) · ZeroBench · ZeroBench_main · avg@5 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

43% · Claude Opus 5.5 (max, with fallback) · ZeroBench · ZeroBench_main · pass@5 · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

OfficeQA Pro ↗62.2%

Claude Opus 5.5 (max, with fallback) · OfficeQA Pro · OfficeQA Pro v2 ↗

HANDBOOK.md ↗41.5%

Claude Opus 5.5 (max, with fallback) · HANDBOOK.md · Kimi K3.1 Pro · complete release comparison ↗

DECK-Arena ↗1253 Elo

Claude Opus 5.5 (max, with fallback) · DECK-Arena · DECK-Arena (Internal, Elo-score) · Kimi K3.1 Pro · complete release comparison ↗

Design Bench ↗927 Elo

Claude Opus 5.5 (max, with fallback) · Design Bench · Design Bench (Internal, Elo-score) · Kimi K3.1 Pro · complete release comparison ↗

Kimi Hallucination Benchmark ↗15%

Claude Opus 5.5 (max, with fallback) · Kimi Hallucination Benchmark · Kimi Hallu Bench Hard w/ tools (Internal, lower is better) · Kimi K3.1 Pro · complete release comparison ↗

1 more reported result

6.8% · Claude Opus 5.5 (max, with fallback) · Kimi Hallucination Benchmark · Kimi Hallu Bench w/o tools (Internal, lower is better) · Kimi K3.1 Pro · complete release comparison · Introducing Kimi K3.1 Pro ↗

ExploitBench ↗91%

Claude Opus 5.5 (max, with fallback) · ExploitBench · Kimi K3.1 Pro · complete release comparison ↗

ExploitGym ↗34.5%

Claude Opus 5.5 (max, with fallback) · ExploitGym · Kimi K3.1 Pro · complete release comparison ↗

Vending-Bench 2 ↗9235.24833 USD

Claude Opus 5.5 · End-of-year balance ↗

GDPx v1.0 ↗12.0%

Claude Opus 5.5 · GDPx v1.0 · Overall · intended grading ↗

53 more reported results

11.7% · Claude Opus 5.5 · GDPx v1.0 · Eight-domain subset · intended grading · GDPx v1.0 — SemiAnalysis ↗

17.6% · Claude Opus 5.5 · GDPx v1.0 · Recursive Self-Improvement · intended grading · GDPx v1.0 — SemiAnalysis ↗

4.8% · Claude Opus 5.5 · GDPx v1.0 · Chip Design · intended grading · GDPx v1.0 — SemiAnalysis ↗

9.3% · Claude Opus 5.5 · GDPx v1.0 · Cybersecurity · intended grading · GDPx v1.0 — SemiAnalysis ↗

20.5% · Claude Opus 5.5 · GDPx v1.0 · Biology · intended grading · GDPx v1.0 — SemiAnalysis ↗

0.0% · Claude Opus 5.5 · GDPx v1.0 · Finance · intended grading · GDPx v1.0 — SemiAnalysis ↗

0.0% · Claude Opus 5.5 · GDPx v1.0 · Legal · intended grading · GDPx v1.0 — SemiAnalysis ↗

27.4% · Claude Opus 5.5 · GDPx v1.0 · Quant Trading · intended grading · GDPx v1.0 — SemiAnalysis ↗

18.7% · Claude Opus 5.5 · GDPx v1.0 · Hardware Design · intended grading · GDPx v1.0 — SemiAnalysis ↗

16.5% · Claude Opus 5.5 · GDPx v1.0 · Materials Science · intended grading · GDPx v1.0 — SemiAnalysis ↗

8.7% · Claude Opus 5.5 · GDPx v1.0 · Alignment · intended grading · GDPx v1.0 — SemiAnalysis ↗

8.9% · Claude Opus 5.5 · GDPx v1.0 · Software Engineering · intended grading · GDPx v1.0 — SemiAnalysis ↗

22.5% · Claude Opus 5.5 · GDPx v1.0 · Overall · partial credit · GDPx v1.0 — SemiAnalysis ↗

24.9% · Claude Opus 5.5 · GDPx v1.0 · Eight-domain subset · partial credit · GDPx v1.0 — SemiAnalysis ↗

17.6% · Claude Opus 5.5 · GDPx v1.0 · Recursive Self-Improvement · partial credit · GDPx v1.0 — SemiAnalysis ↗

12.1% · Claude Opus 5.5 · GDPx v1.0 · Chip Design · partial credit · GDPx v1.0 — SemiAnalysis ↗

11.8% · Claude Opus 5.5 · GDPx v1.0 · Cybersecurity · partial credit · GDPx v1.0 — SemiAnalysis ↗

27.4% · Claude Opus 5.5 · GDPx v1.0 · Biology · partial credit · GDPx v1.0 — SemiAnalysis ↗

54.6% · Claude Opus 5.5 · GDPx v1.0 · Finance · partial credit · GDPx v1.0 — SemiAnalysis ↗

29.9% · Claude Opus 5.5 · GDPx v1.0 · Legal · partial credit · GDPx v1.0 — SemiAnalysis ↗

27.4% · Claude Opus 5.5 · GDPx v1.0 · Quant Trading · partial credit · GDPx v1.0 — SemiAnalysis ↗

18.7% · Claude Opus 5.5 · GDPx v1.0 · Hardware Design · partial credit · GDPx v1.0 — SemiAnalysis ↗

30.1% · Claude Opus 5.5 · GDPx v1.0 · Materials Science · partial credit · GDPx v1.0 — SemiAnalysis ↗

8.7% · Claude Opus 5.5 · GDPx v1.0 · Alignment · partial credit · GDPx v1.0 — SemiAnalysis ↗

8.9% · Claude Opus 5.5 · GDPx v1.0 · Software Engineering · partial credit · GDPx v1.0 — SemiAnalysis ↗

8.7% · Claude Opus 5.5 · GDPx v1.0 · Aria · Chip Design · intended grading · GDPx v1.0 — SemiAnalysis ↗

0.0% · Claude Opus 5.5 · GDPx v1.0 · Arcverex · Chip Design · intended grading · GDPx v1.0 — SemiAnalysis ↗

5.7% · Claude Opus 5.5 · GDPx v1.0 · Ergodex · Chip Design · intended grading · GDPx v1.0 — SemiAnalysis ↗

9.3% · Claude Opus 5.5 · GDPx v1.0 · Arithmetic · Cybersecurity · intended grading · GDPx v1.0 — SemiAnalysis ↗

20.5% · Claude Opus 5.5 · GDPx v1.0 · LatchBio · Biology · intended grading · GDPx v1.0 — SemiAnalysis ↗

0.0% · Claude Opus 5.5 · GDPx v1.0 · Living Context · Finance · intended grading · GDPx v1.0 — SemiAnalysis ↗

0.0% · Claude Opus 5.5 · GDPx v1.0 · Chakra · Legal · intended grading · GDPx v1.0 — SemiAnalysis ↗

27.4% · Claude Opus 5.5 · GDPx v1.0 · Third Moment · Quant Trading · intended grading · GDPx v1.0 — SemiAnalysis ↗

17.6% · Claude Opus 5.5 · GDPx v1.0 · Extrapolate · Recursive Self-Improvement · intended grading · GDPx v1.0 — SemiAnalysis ↗

18.7% · Claude Opus 5.5 · GDPx v1.0 · Beautiful Machines · Hardware Design · intended grading · GDPx v1.0 — SemiAnalysis ↗

16.7% · Claude Opus 5.5 · GDPx v1.0 · ValenceForge · Materials Science · intended grading · GDPx v1.0 — SemiAnalysis ↗

16.4% · Claude Opus 5.5 · GDPx v1.0 · Emulated · Materials Science · intended grading · GDPx v1.0 — SemiAnalysis ↗

8.7% · Claude Opus 5.5 · GDPx v1.0 · Fleet · Alignment · intended grading · GDPx v1.0 — SemiAnalysis ↗

8.9% · Claude Opus 5.5 · GDPx v1.0 · Datacurve · Software Engineering · intended grading · GDPx v1.0 — SemiAnalysis ↗

30.6% · Claude Opus 5.5 · GDPx v1.0 · Aria · Chip Design · partial credit · GDPx v1.0 — SemiAnalysis ↗

0.0% · Claude Opus 5.5 · GDPx v1.0 · Arcverex · Chip Design · partial credit · GDPx v1.0 — SemiAnalysis ↗

5.7% · Claude Opus 5.5 · GDPx v1.0 · Ergodex · Chip Design · partial credit · GDPx v1.0 — SemiAnalysis ↗

11.8% · Claude Opus 5.5 · GDPx v1.0 · Arithmetic · Cybersecurity · partial credit · GDPx v1.0 — SemiAnalysis ↗

27.4% · Claude Opus 5.5 · GDPx v1.0 · LatchBio · Biology · partial credit · GDPx v1.0 — SemiAnalysis ↗

54.6% · Claude Opus 5.5 · GDPx v1.0 · Living Context · Finance · partial credit · GDPx v1.0 — SemiAnalysis ↗

29.9% · Claude Opus 5.5 · GDPx v1.0 · Chakra · Legal · partial credit · GDPx v1.0 — SemiAnalysis ↗

27.4% · Claude Opus 5.5 · GDPx v1.0 · Third Moment · Quant Trading · partial credit · GDPx v1.0 — SemiAnalysis ↗

17.6% · Claude Opus 5.5 · GDPx v1.0 · Extrapolate · Recursive Self-Improvement · partial credit · GDPx v1.0 — SemiAnalysis ↗

18.7% · Claude Opus 5.5 · GDPx v1.0 · Beautiful Machines · Hardware Design · partial credit · GDPx v1.0 — SemiAnalysis ↗

38.7% · Claude Opus 5.5 · GDPx v1.0 · ValenceForge · Materials Science · partial credit · GDPx v1.0 — SemiAnalysis ↗

21.4% · Claude Opus 5.5 · GDPx v1.0 · Emulated · Materials Science · partial credit · GDPx v1.0 — SemiAnalysis ↗

8.7% · Claude Opus 5.5 · GDPx v1.0 · Fleet · Alignment · partial credit · GDPx v1.0 — SemiAnalysis ↗

8.9% · Claude Opus 5.5 · GDPx v1.0 · Datacurve · Software Engineering · partial credit · GDPx v1.0 — SemiAnalysis ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

Claude Sonnet 5.5 →Anthropic

11.6 percentile-point gap

59 of 74 shared benchmark comparisons

Claude Fable 5.1 →Anthropic

13.2 percentile-point gap

63 of 74 shared benchmark comparisons

GPT-6 Astra →OpenAI

17.0 percentile-point gap

69 of 74 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

POST-TRAINING / SOURCE DISCLOSURE

What the Claude Opus 5.5 report says

Anthropic describes fine-tuning after pretraining to align the assistant with Claude’s constitution. The system card also documents human preference selection, safety evaluation and adversarial testing.

Post-training and fine-tuning guided by Claude’s constitution
Human preference and safety-evaluation data

Constitutional AI · background →Preference learning · background →

The system card does not disclose a production PPO, DPO or GRPO recipe. Constitutional AI is a related method note; its particular SL-CAI/RL-CAI implementation is not asserted for this release.

Claude Opus 5.5 System Card ↗
Anthropic · 2026

Filter comparisons to this model →

Reports featuring this model

Kimi K3.1 Pro · Kimi →

Claude Opus 5.5 · Anthropic →

Muse Wonder and Muse Wonder Full · Meta →

Gemini 4 Argon · Google →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

ARC Prize - Leaderboard ↗
ARC Prize Foundation
Arena leaderboard — WebDev evaluation ↗
Arena
AA-3D: Interactive 3D Experience Benchmark | Artificial Analysis ↗
Artificial Analysis
AA-AnalystAgent Benchmark | Artificial Analysis ↗
Artificial Analysis
AA-Briefcase v1.1: Agentic Knowledge Work Benchmark | Artificial Analysis ↗
Artificial Analysis
AA-PersonalityBench: AI Personality & Communication Style Benchmark | Artificial Analysis ↗
Artificial Analysis
Artificial Analysis Intelligence Index v5.0 — Artificial Analysis ↗
Artificial Analysis
AA-LCR v1.1: Long Context Reasoning Benchmark | Artificial Analysis ↗
Artificial Analysis
AutomationBench-AA: Agentic SaaS Workflow Benchmark | Artificial Analysis ↗
Artificial Analysis
CritPt Benchmark | Artificial Analysis ↗
Artificial Analysis
CritPt v2 Benchmark | Artificial Analysis ↗
Artificial Analysis
CritPt v2 — Hard Subset — Artificial Analysis ↗
Artificial Analysis
FrontierCode 1.1 Benchmark | Artificial Analysis ↗
Artificial Analysis
FrontierCode 2.0 Benchmark — Artificial Analysis ↗
Artificial Analysis
GDP.pdf Benchmark | Artificial Analysis ↗
Artificial Analysis
GDPval-AA v2.1 Benchmark | Artificial Analysis ↗
Artificial Analysis
Harvey LAB-AA v1.1 Benchmark | Artificial Analysis ↗
Artificial Analysis
HealthBench Professional Benchmark | Artificial Analysis ↗
Artificial Analysis
HLE-Diamond Benchmark — Artificial Analysis ↗
Artificial Analysis
Humanity's Last Exam Benchmark | Artificial Analysis ↗
Artificial Analysis
ITBench-AA Benchmark | Artificial Analysis ↗
Artificial Analysis
MLCR-AA: Medical Long Context Reasoning Benchmark | Artificial Analysis ↗
Artificial Analysis
MMMU-Pro Benchmark | Artificial Analysis ↗
Artificial Analysis
AA-Omniscience v1.1: Knowledge and Hallucination Benchmark — Artificial Analysis ↗
Artificial Analysis
SciCode Benchmark | Artificial Analysis ↗
Artificial Analysis
Terminal-Bench-Science 0.1 Benchmark | Artificial Analysis ↗
Artificial Analysis
Terminal-Bench 4.0 Benchmark | Artificial Analysis ↗
Artificial Analysis
TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts — published result table ↗
Jaffe, Oliver, Sherburn, Dane · 2026
FrontierCode Leaderboard ↗
Cognition
DTBench — Conceptual Reasoning Index ↗
Conceptual Reasoning Index
LMCA — Conceptual Reasoning Index ↗
Conceptual Reasoning Index
Cursor · CursorBench ↗
Cursor
Gemini 4 Argon Model Card ↗
Google DeepMind · 2026
Muse Wonder | Meta ↗
Meta AI
APEX-Agents ↗
Epoch AI
ARC-AGI-2 ↗
Epoch AI
Chess Puzzles ↗
Epoch AI
CritPt v1 ↗
Epoch AI
Earthborne Rangers (EBR-bench) ↗
Epoch AI
FrontierMath Erdős ↗
Epoch AI
FrontierMath Tier 4 (v2) ↗
Epoch AI
FrontierMath Tiers 1-3 (v2) ↗
Epoch AI
Furniture Assembly Benchmark ↗
Epoch AI
GPQA Diamond ↗
Epoch AI
MirrorCode ↗
Epoch AI
Mystery Game Puzzles ↗
Epoch AI
OTIS Mock AIME 2024-2025 ↗
Epoch AI
SimpleQA Verified ↗
Epoch AI
SpireClimb ↗
Epoch AI
Vending Bench 2 — methodology and published results ↗
Epoch AI · 2026
AI benchmarks & capabilities ↗
Epoch AI
GDPx v1.0: The General Agentic Benchmark — SemiAnalysis ↗
SemiAnalysis · 2026
WeirdML ↗
Håvard Tveit Ihle
DistressBench — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
Scale Labs — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
HiL-Bench (Human-in-Loop Benchmark) — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
Scale Labs — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
Humanity's Sixth Sense — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
Scale Labs — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
Scale Labs — evaluation methodology and results ↗
Scale Labs / SEAL · 2026
Introducing Muse Wonder ↗
Meta AI · 2026
GDP.pdf Benchmark | Surge AI ↗
Surge AI
Introducing Claude Opus 5.5 ↗
Anthropic
FrontierSWE — software engineering benchmark ↗
FrontierSWE
Introducing Kimi K3.1 Pro ↗
Moonshot AI · 2026
ErdősBench: research mathematics, scoped progress and proof audits ↗
Ulam AI · 2026
STEMBench — Six-Domain Model Evaluation | Ulam ↗
Ulam
BioMysteryBench — evaluation methodology and results ↗
Vals AI · 2026
Code Migration v1 — evaluation methodology and results ↗
Vals AI · 2026
Code Migration — evaluation methodology and results ↗
Vals AI · 2026
CUA-bench — evaluation methodology and results ↗
Vals AI · 2026
CyberBench — evaluation methodology and results ↗
Vals AI · 2026
Excel Modeling Benchmark v1 — evaluation methodology and results ↗
Vals AI · 2026
Excel Modeling Benchmark — evaluation methodology and results ↗
Vals AI · 2026
Finance Agent v2 — evaluation methodology and results ↗
Vals AI · 2026
Finance Agent — evaluation methodology and results ↗
Vals AI · 2026
Vals Finance Index — evaluation methodology and results ↗
Vals AI · 2026
HLAB-Vals-1.1 — evaluation methodology and results ↗
Vals AI · 2026
Harvey's Legal Agent Benchmark v1 — evaluation methodology and results ↗
Vals AI · 2026
IOI — evaluation methodology and results ↗
Vals AI · 2026
Vals Legal Index — evaluation methodology and results ↗
Vals AI · 2026
Legal Research Bench — evaluation methodology and results ↗
Vals AI · 2026
MedCode — evaluation methodology and results ↗
Vals AI · 2026
MedScribe — evaluation methodology and results ↗
Vals AI · 2026
SAFE-Teen: Teen Mental Health & Safety | Vals AI A conversation branches after the chatbot turns down a request — evaluation methodology and results ↗
Vals AI · 2026
MysteryMechanism — evaluation methodology and results ↗
Vals AI · 2026
ProgramBench — evaluation methodology and results ↗
Vals AI · 2026
ProofBench — evaluation methodology and results ↗
Vals AI · 2026
Public Benefits Bench — evaluation methodology and results ↗
Vals AI · 2026
Robo-bench — evaluation methodology and results ↗
Vals AI · 2026
Vals RSI Index — evaluation methodology and results ↗
Vals AI · 2026
SAGE — evaluation methodology and results ↗
Vals AI · 2026
Vals Science Index — evaluation methodology and results ↗
Vals AI · 2026
SRE Bench — evaluation methodology and results ↗
Vals AI · 2026
SWE-bench Multimodal — evaluation methodology and results ↗
Vals AI · 2026
Tax Agent Bench — evaluation methodology and results ↗
Vals AI · 2026
Terminal-Bench 4.0 — evaluation methodology and results ↗
Vals AI · 2026
Terminal-Bench-Science — evaluation methodology and results ↗
Vals AI · 2026
Time Horizon Index: KSP — evaluation methodology and results ↗
Vals AI · 2026
Vals Index — evaluation methodology and results ↗
Vals AI · 2026
Vibe Code Bench 1-100 — evaluation methodology and results ↗
Vals AI · 2026
Vibe Code Bench — evaluation methodology and results ↗
Vals AI · 2026

Names used in source tables

Claude Opus 5.5 · Claude Opus 5.5 (Max, Default Fallback) · Claude Opus 5.5 (Mini-SWE-agent) · Claude Opus 5.5 (OpenHands) · Claude Opus 5.5 (high) · Claude Opus 5.5 (low) · Claude Opus 5.5 (max) · Claude Opus 5.5 (max, with fallback) · Claude Opus 5.5 (medium) · Claude Opus 5.5 (xhigh) · Claude Opus 5.5 xhigh · Claude-Opus-5.5 (xhigh) · Opus 5.5 · Opus 5.5 (max) · claude-opus-5-5 · claude-opus-5.5