AIEvals

Models

Claude Sonnet 5.5

Anthropic’s Claude 5.5 model for coding and knowledge work, with adaptive reasoning and configurable effort.

AnthropicReleased 28 September 2026Context 1M tokensOutput 128k tokensPrivate weights$2.00 in / $10.00 out per Mtok

Family Claude 5.559 benchmarks with a published result6 independent first places among the 46 models this site listsRead 17 October 2026

Where it stands

Rank compares tracked models with a score. Published shows the source’s rank and field size, when available.

BenchmarkScoreRankPublishedLeaderBehind
Vals IndexVals AI67.04%2of 332of 41Gemini 4 Argon1.86 pts
Artificial Analysis Intelligence IndexArtificial Analysis562of 42no list readClaude Opus 5.52 pts
Terminal-Bench 4.0Artificial Analysis63.64%1of 40no list readClaude Sonnet 5.5—
Vibe Code Bench v1.1Vals AI92.39%1of 381of 106Claude Sonnet 5.5—
Terminal-Bench Science 0.1Artificial Analysis53.33%4of 24no list readGPT-6 Astra10.00 pts
MLCR-AAArtificial Analysis75.00%1of 37no list readClaude Sonnet 5.5—
Harvey LAB-AAArtificial Analysis93.07%6of 20no list readMuse Spark 1.3 Max2.41 pts
AutomationBench-AAArtificial Analysis71.33%2of 40no list readGemini 4 Argon6.18 pts
SciCodeArtificial Analysis61.00%4of 40no list readClaude Opus 5.55.90 pts
GDP.pdfArtificial Analysis25.80%7of 40no list readGPT-6.1 Sol5.20 pts
CritPtArtificial Analysis31.43%5of 42no list readGPT-5.6 Sol0.86 pts
AA-OmniscienceArtificial Analysis32.38of 42no list readClaude Opus 5.514.12
AA-LCR v1.1Artificial Analysis82.67%16of 42no list readKimi K36.00 pts
Terminal-Bench 2.1Vals AI83.15%6of 356of 75Claude Opus 5.54.49 pts
ProofBench v1.1Vals AI100.00%1of 341of 46Claude Sonnet 5.5—
Harvey's Legal Agent BenchmarkVals AI2.92%29of 3738of 73Muse Spark 1.222.50 pts
MysteryMechanismVals AI49.10%3of 213of 21GPT-6 Astra4.05 pts
BioMysteryBenchVals AI81.11%1of 191of 21Claude Sonnet 5.5—
Code MigrationVals AI69.83%1of 371of 71Claude Sonnet 5.5—
CyberBench v1.1Vals AI59.58%25of 3132of 43GPT-6 Sol18.40 pts
EMBVals AI75.71%3of 373of 69Claude Fable 5.10.96 pts
Finance Agent (v2)Vals AI58.10%9of 3710of 73Gemini 4 Argon7.30 pts
IOIVals AI83.06%9of 319of 38Gemini 4 Argon16.94 pts
Legal Research BenchVals AI48.08%8of 378of 72Muse Spark 1.3 Max7.21 pts
MedCodeVals AI52.92%9of 3512of 104Claude Opus 510.65 pts
MedScribeVals AI91.10%3of 363of 106Claude Opus 5.50.33 pts
Public Benefits Bench v1.1Vals AI67.19%13of 2913of 45Claude Opus 59.74 pts
SAGEVals AI51.77%7of 3211of 90Claude Opus 4.74.33 pts
SRE BenchVals AI30.15%6of 156of 16GPT-6 Astra26.72 pts
Tax Agent BenchVals AI73.39%4of 374of 64Claude Fable 5.14.25 pts
GDPval-AA v2.1Artificial Analysis1,844.212of 42no list readClaude Opus 5.51.96
AA-Briefcase v1.1Artificial Analysis1,810.92of 41no list readClaude Opus 5.510.95
Humanity's Last ExamArtificial Analysis54.96%5of 42no list readClaude Opus 5.56.39 pts

Nearby on Vals Index

The two models above and below on the Vals Index (Vals AI), on every column where this model has a published number.

ModelVals Index%Artificial Analysis Intelligence IndexindexTerminal-Bench 4.0%Vibe Code Bench v1.1%Terminal-Bench Science 0.1%
Gemini 4 ArgonGoogle68.90%5357.07%91.91%–
Claude Sonnet 5.5Anthropic67.04%5663.64%92.39%53.33%
Claude Opus 5.5Anthropic66.97%5859.60%90.29%59.05%
Claude Fable 5.1Anthropic65.83%5352.02%90.26%43.33%
Bold marks a figure better than this model’s on the same column, whichever model it belongs to. Ranks cover the models shown from the stated source.

Cost and speed, as published

Additional measurements from Artificial Analysis’ own row. Read 30 September 2026.

As publishedCost per taskOutput speed
Claude Sonnet 5.5 (Adaptive Reasoning, Max Effort, Default Fallback)$7.60138.7 tok/s

Results by benchmark

Independent and vendor results retain their own harnesses and effort settings.

BenchmarkIndependent runAs reported by a publisher
SWE-Bench Pro (version unspecified)Repository codingno published result
81.3%SWE-Bench Pro (version unspecified)Anthropic · 7 October 2026Rank 1 of 2 in this publisher table
SWE-bench MultimodalRepository codingno published result
54.3%SWE-bench MultimodalAnthropic · 7 October 2026Rank 1 of 3 in this publisher table
FrontierCode v1.1 (Extended)Agentic codingno published result
59.1%FrontierCode v1.1 (Extended)Anthropic · 7 October 2026Rank 1 of 2 in this publisher table
64.4%FrontierCode v1.1 (Extended, xhigh)Anthropic · 7 October 2026Rank 1 of 1 in this publisher table
ProgramBench (hidden test pass rate)Long-context codingno published result
79.7%ProgramBench (hidden test pass rate)Anthropic · 7 October 2026Rank 3 of 3 in this publisher table
DRACO (Anthropic harness)Deep researchno published result
87.0%DRACO (Anthropic harness)Anthropic · 7 October 2026Rank 2 of 4 in this publisher table
WANDR (frozen index)Wide researchno published result
70.0%WANDR (frozen index)Anthropic · 7 October 2026Rank 2 of 4 in this publisher table
BenchCAD Vision2Code (no tools)CAD reasoningno published result
0.747BenchCAD Vision2Code (no tools)Anthropic · 7 October 2026Rank 1 of 5 in this publisher table
BenchCAD Vision2Code (with tools)CAD reasoningno published result
0.963BenchCAD Vision2Code (with tools)Anthropic · 7 October 2026Rank 1 of 5 in this publisher table
OSWorld 2.1 (offline strict pass)Computer useno published result
48.8%OSWorld 2.1 (offline strict pass)Anthropic · 7 October 2026Rank 2 of 6 in this publisher table
OfficeQADocument reasoningno published result
76.9%OfficeQAAnthropic · 7 October 2026Rank 1 of 3 in this publisher table
OfficeQA ProDocument reasoningno published result
65.6%OfficeQA ProAnthropic · 7 October 2026Rank 1 of 3 in this publisher table
HealthBench Professional (length-adjusted)Healthcareno published result
69.2%HealthBench Professional (length-adjusted)Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
PhysicianBenchHealthcareno published result
63.2%PhysicianBenchAnthropic · 7 October 2026Rank 2 of 6 in this publisher table
Global MMLUMultilingual knowledgeno published result
92.1%Global MMLUAnthropic · 7 October 2026Rank 3 of 4 in this publisher table
MILUMultilingual knowledgeno published result
91.6%MILUAnthropic · 7 October 2026Rank 3 of 4 in this publisher table
Morphology-to-molecule matchingLife sciencesno published result
25.0%Morphology-to-molecule matchingAnthropic · 7 October 2026Rank 2 of 2 in this publisher table
Medicinal chemistry (ADME)Life sciencesno published result
65.3%Medicinal chemistry (ADME)Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
Protein design (sequence generation)Life sciencesno published result
51.0%Protein design (sequence generation)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
Protein design (library ranking)Life sciencesno published result
54.8%Protein design (library ranking)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
Protocols (troubleshooting)Life sciencesno published result
67.3%Protocols (troubleshooting)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
Protocols (understanding V2)Life sciencesno published result
66.6%Protocols (understanding V2)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
MLCR-AAMedical reasoning75.00%Artificial Analysisnot published
Harvey LAB-AALegal reasoning93.07%Artificial Analysisnot published
SWE-bench MultilingualRepository codingno published result
90.3%SWEBench MultilingualAnthropic · 7 October 2026Rank 1 of 3 in this publisher table
AutomationBench-AAWorkflow automation71.33%Artificial Analysisnot published
SciCodeScientific coding61.00%Artificial Analysisnot published
GDP.pdfDocument reasoning25.80%Artificial Analysisnot published
CritPtScientific reasoning31.43%Artificial Analysisnot published
AA-OmniscienceFactual knowledge32.3Artificial Analysisnot published
AA-LCR v1.1Long-context reasoning82.67%Artificial Analysisnot published
Terminal-Bench 4.0Agentic coding63.64%Artificial Analysis
64.14%Vals AI
70.6%Terminal-Bench 4.0Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
70.6%Terminal-Bench 4.0Anthropic · 7 October 2026Rank 1 of 6 in this publisher table
70.6%Terminal-Bench 4.0Anthropic · 28 September 2026Rank 1 of 3 in this publisher table
Terminal-Bench 2.1Agentic coding83.15%Vals AInot published
Vibe Code Bench v1.1Product coding92.39%Vals AInot published
Terminal-Bench Science 0.1Agentic science53.33%Artificial Analysis
38.57%Vals AI
59.9%Terminal-Bench-Science 0.1Anthropic · 7 October 2026Rank 1 of 5 in this publisher table
ProofBench v1.1Mathematics100.00%Vals AInot published
Harvey's Legal Agent BenchmarkProfessional knowledge2.92%Vals AInot published
MysteryMechanismScientific reasoning49.10%Vals AInot published
Vals IndexComposite67.04%Vals AInot published
BioMysteryBenchAgentic science81.11%Vals AInot published
Code MigrationAgentic coding69.83%Vals AInot published
CyberBench v1.1Security59.58%Vals AInot published
EMBFinancial modelling75.71%Vals AInot published
Finance Agent (v2)Financial analysis58.10%Vals AInot published
IOICompetitive programming83.06%Vals AInot published
Legal Research BenchProfessional knowledge48.08%Vals AInot published
MedCodeHealthcare administration52.92%Vals AInot published
MedScribeHealthcare administration91.10%Vals AInot published
Public Benefits Bench v1.1Public services67.19%Vals AInot published
SAGEEducation51.77%Vals AInot published
SRE BenchSecurity30.15%Vals AInot published
Tax Agent BenchProfessional knowledge73.39%Vals AInot published
OSWorld 2.1Computer useno published result
83.9%OSWorld 2.1 (offline partial)Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
83.9%OSWorld 2.1 (offline partial)Anthropic · 7 October 2026Rank 2 of 6 in this publisher table
80.1%OSWorld 2.1 (partial)Anthropic · 28 September 2026Rank 2 of 3 in this publisher table
CursorBench 4.0Agentic codingno published result
55.5%CursorBench 4.0Anthropic · 28 September 2026Rank 2 of 3 in this publisher table
FrontierCode v1.1Agentic codingno published result
52.1%FrontierCode v1.1 (Main)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
46.2%FrontierCode v1.1 (Main, max)Anthropic · 7 October 2026Rank 2 of 3 in this publisher table
52.1%FrontierCode v1.1 (Main, xhigh)Anthropic · 7 October 2026Rank 1 of 2 in this publisher table
46.2%FrontierCode 1.1 (Main)Anthropic · 28 September 2026Rank 3 of 4 in this publisher table
GDPval-AA v2.1Knowledge work1,844.21Artificial Analysis
1,840GDPval-AA v2.1 (max)Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
1,840GDPval-AA v2.1 (max)Anthropic · 7 October 2026Rank 1 of 5 in this publisher table
1,844GDPval-AA v2.1Anthropic · 28 September 2026Rank 2 of 4 in this publisher table
AA-Briefcase v1.1Knowledge work1,810.9Artificial Analysis
1,824AA-Briefcase v1.1 (max)Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
1,824AA-Briefcase v1.1 (max)Anthropic · 7 October 2026Rank 1 of 5 in this publisher table
1,811AA-Briefcase v1.1Anthropic · 28 September 2026Rank 2 of 4 in this publisher table
ChartographyVisual chart recognitionno published result
61.6%Chartography (no tools)Anthropic · 7 October 2026Rank 1 of 4 in this publisher table
61.6%Chartography (no tools)Anthropic · 7 October 2026Rank 3 of 10 in this publisher table
90.2%Chartography (with tools)Anthropic · 7 October 2026Rank 1 of 5 in this publisher table
61.6%Chartography (no tools)Anthropic · 28 September 2026Rank 2 of 4 in this publisher table
Humanity's Last ExamMultidisciplinary reasoning54.96%Artificial Analysis
56.9%Humanity's Last Exam (no tools)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
64.5%Humanity's Last Exam (with tools)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
56.9%Humanity's Last Exam (no tools)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
64.5%Humanity's Last Exam (with tools)Anthropic · 7 October 2026Rank 1 of 3 in this publisher table
64.5%Humanity's Last Exam (with tools)Anthropic · 28 September 2026Rank 2 of 3 in this publisher table
Artificial Analysis Intelligence IndexComposite56Artificial Analysisnot published
Independent column: the suite named under each figure, one harness per suite. Publisher column: each lab’s own table, reproduced on the benchmark’s page. Publisher ranks compare reported scores in the same table row only; ties share a rank and unreported cells are excluded.

Sources

Read on 17 October 2026.