Skip to main content

Model profile

Claude Sonnet 4.6

AnthropicSupersededReleased February 2026
Data verified
Superseded:Anthropic has released newer models in this line —Claude Sonnet 5
Overall Score
65.07Public #32 of 200Verified #27 of 99
Arena Elo
1471
Eligible category ranks
3of 8
Price (1M tokens)
$3 in / $15 out
API pricing
Speed
44tok/s
Context
200K

Evidence coverage

33 of 321 tracked benchmarks are published. 18 are verified and 15 provisional. 7 of 8 categories are measured.

Updated July 20, 2026Methodology
Published / tracked
33 / 321
Verified
18
Provisional
15
Categories with evidence
7 / 8

Evidence by category

  • Agentic8 benchmarks
    Mixed evidence
  • Coding7 benchmarks
    Mixed evidence
  • Reasoning2 benchmarks
    Reported
  • Knowledge10 benchmarks
    Mixed evidence
  • Math2 benchmarks
    Verified
  • Multilingual0 benchmarks
    Not measured
  • Multimodal3 benchmarks
    Mixed evidence
  • Inst. Following1 benchmark
    Reported
ProprietaryNon-Reasoning
Confidence:
High
base

Claude Sonnet 4.6 ranks #32 out of 200 models on the public leaderboard with an overall score of 65.07/100. It also ranks #27 out of 99 on the verified leaderboard. While not a frontier model, it offers specific advantages depending on the use case.

Claude Sonnet 4.6 is a proprietary model with a 200K token context window. It processes queries without explicit chain-of-thought reasoning, offering faster response times and lower token usage.

This profile currently has 33 of 321 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Knowledge (#15), while its weakest is Agentic (#42). This performance profile makes it particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.

Peer position

Exact provisional scores and ranks for the closest listed peers. A score can appear before a model clears the evidence threshold for a rank, so equal scores can have different rank states.

Range 64.1865.32

  1. Claude Sonnet 5
    Anthropic
    #2965.32
    Claude Sonnet 5 is #29 with a score of 65.32.
    Compare
  2. Qwen3.6 Plus
    Alibaba
    #3065.2
    Qwen3.6 Plus is #30 with a score of 65.2.
    Compare
  3. Grok 4.3
    xAI
    #3165.1
    Grok 4.3 is #31 with a score of 65.1.
    Compare
  4. Claude Sonnet 4.6Current model
    Anthropic
    #3265.07
    Claude Sonnet 4.6 is #32 with a score of 65.07.
  5. Gemini 3.5 Flash
    Google
    #3364.75
    Gemini 3.5 Flash is #33 with a score of 64.75.
    Compare
  6. Claude Opus 4.5
    Anthropic
    #3464.22
    Claude Opus 4.5 is #34 with a score of 64.22.
    Compare
  7. Claude Opus 4.6 (Adaptive)
    Anthropic
    #3564.18
    Claude Opus 4.6 (Adaptive) is #35 with a score of 64.18.
    Compare

Category percentile

More

Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.

  1. Knowledge73%
    Eligible cohort rank #15 of 52Category score 79.7
  2. Coding82%
    Eligible cohort rank #23 of 122Category score 59.6
  3. Agentic65%
    Eligible cohort rank #42 of 119Category score 51.2

Category evidence

Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #42 of 119Percentile 65thWeight 22%8 benchmarksMixed sources51.2
CodingRank #23 of 122Percentile 82ndWeight 20%7 benchmarksMixed sources59.6
ReasoningRank Not rankedWeight 17%2 benchmarksReported74.3
KnowledgeRank #15 of 52Percentile 73rdWeight 12%10 benchmarksMixed sources79.7
MathRank Not rankedWeight 5%2 benchmarksVerified49.5
MultilingualWeight 7%0 benchmarksNot measuredNot measured
MultimodalRank Not rankedWeight 12%3 benchmarksMixed sources56.5
Inst. FollowingRank Not rankedWeight 5%1 benchmarkReported81.1

Chatbot Arena performance

Scroll horizontally to inspect confidence intervals and vote counts.

Chatbot Arena Elo, confidence interval, and vote count by evaluation view
ViewEloConfidence intervalVotes
Text Overall1471±3.954,975
Coding1528±6.315,132
Math1464±11.32,994
Instruction Following1477±5.817,772
Creative Writing1450±7.49,053
Multi-turn1484±7.19,930
Hard Prompts1504±4.835,237
Hard Prompts (English)1509±6.017,404
Longer Query1494±5.722,658

Benchmark Details

Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.

Agentic8 benchmarks
Terminal-Bench 2.0Provider exact
59.1%Weighted 38%
Source: Anthropic: Claude Sonnet 4.6 system cardProvenance: Provider exact
OSWorld-VerifiedBenchmark exact
72.1%Weighted 34%
Source: OSWorld-Verified leaderboardProvenance: OSWorld-Verified reports the claude-sonnet-4-6 general-model run at 72.11% success rate in the benchmark leaderboard data.
Claw-EvalBenchmark exact
67.8%Display only
Source: Claw-Eval leaderboardProvenance: Claw-Eval reports this model as sonnet46 in the official 2026-05-09 leaderboard snapshot. BenchLM stores the primary Pass^3 value on the local Claw-Eval display key.
CyberGymBenchmark exact
65.2%Display only
Source: CyberGym leaderboardProvenance: CyberGym Level 1 reports Anthropic Agent with Claude Sonnet 4.6 on the public leaderboard. BenchLM stores the reported target-vulnerability reproduction success rate on the local cyberGym key.
τ²-bench resultsReported

τ²-Bench Tool-Agent-User Evaluation

79.5%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Gert LabsBenchmark exact

Gert Labs Composite Game Benchmark

62.92%Display only
Source: Gert Labs rankingsProvenance: Gert Labs reports this composite leaderboard score in the public rankings API. BenchLM scales the source gscore from 0-1 to 0-100 and stores it as a display-only agentic benchmark.
OSWorld 2.0Benchmark exact
8.3%Display only
Source: OSWorld 2.0 paperProvenance: OSWorld 2.0 reports Claude Sonnet 4.6 single action on its 500-step main table. BenchLM stores the binary completion score and notes the corresponding partial score was 41.5%.
JobBenchBenchmark exact
36.9%Display only
Source: JobBench paperProvenance: JobBench reports Sonnet-4.6 under Claude Code on the main-set leaderboard. BenchLM stores the reported main-set score.
Coding7 benchmarks
SWE-RebenchBenchmark exact
60.7%Weighted 20%
Source: SWE-Rebench leaderboardProvenance: Public SWE-Rebench leaderboard lists Claude Sonnet 4.6 at 60.7% resolved rate.
SWE-bench VerifiedProvider exact

Software Engineering Benchmark Verified

79.6%Weighted 16%
Source: Anthropic: Claude Sonnet 4.6 system cardProvenance: Provider exact
React Native EvalsBenchmark exact
80.6%Display only
Source: React Native Evals leaderboardProvenance: React Native Evals reports this exact overall score for Claude Sonnet 4.6 in the public dashboard run finished on 2026-04-28.
Vibe Code BenchBenchmark exact

Vibe Code Bench v1.1

51.48%Display only
Source: Vals AI: Vibe Code Bench v1.1Provenance: Vals Vibe Code Bench v1.1 reports this exact row under anthropic/claude-sonnet-4-6; BenchLM stores it on the local vibeCodeBench key.
cursorBench31Benchmark exact
48.8%Display only
Source: Cursor evals: CursorBench 3.1Provenance: Cursor reports Sonnet 4.6 High at this exact CursorBench 3.1 score on its public evals page. BenchLM stores it on the Claude Sonnet 4.6 row as a display-only coding-agent benchmark.
AA-SciCodeReported

Artificial Analysis SciCode

46.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
FrontierCode 1.1 MainBenchmark exact
24.3%Display only
Source: Cognition: FrontierCode 1.1Provenance: Cognition reports Claude Sonnet 4.6 at 24.3% on FrontierCode 1.1 Main, using the best max effort row from the published data JSON.
Reasoning2 benchmarks
AA-LCRReported

Artificial Analysis Long Context Reasoning

57.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
CritPtReported

Critical Physics Tasks

0.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Knowledge10 benchmarks
HLEProvider exact

Humanity's Last Exam

49%Weighted 45%
Source: Anthropic: Claude Sonnet 4.6 system cardProvenance: Anthropic Table 2.1.A reports HLE with tools at 49.0% for Claude Sonnet 4.6.
MMLU-ProSecondary exact

Massive Multitask Language Understanding Professional

79.2%Weighted 30%
Source: Qwen3.6-Plus comparison tableProvenance: Secondary exact
GPQAProvider exact

Graduate-Level Google-Proof Q&A

89.9%Weighted 7%
Source: Anthropic: Claude Sonnet 4.6 system cardProvenance: Provider exact
SuperGPQASecondary exact

SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines

95%Weighted 7%
Source: Qwen3.6-Plus comparison tableProvenance: Secondary exact
Artificial Analysis Intelligence IndexReported
35.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-GPQA DiamondReported

Artificial Analysis GPQA Diamond

79.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-HLEReported

Artificial Analysis Humanity's Last Exam

13.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience IndexReported

Artificial Analysis Omniscience Index

-2.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience AccuracyReported

Artificial Analysis Omniscience Accuracy

38.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience Hallucination RateReported

Artificial Analysis Omniscience Hallucination Rate

65.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Math2 benchmarks
FrontierMath v2 (Tiers 1-3)Benchmark exact

FrontierMath v2 Tiers 1-3

32.400%Weighted 30%
Source: Epoch AI FrontierMath v2 leaderboardProvenance: Epoch AI reports FrontierMath v2 Tiers 1-3 at 32.4% for claude-sonnet-4-6_16K. BenchLM selects the highest published thinking effort for the model and stores the v2 benchmark slice separately.
FrontierMath v2 (Tier 4)Benchmark exact

FrontierMath v2 Tier 4

8.300%Weighted 10%
Source: Epoch AI FrontierMath v2 leaderboardProvenance: Epoch AI reports FrontierMath v2 Tier 4 at 8.3% for claude-sonnet-4-6_16K. BenchLM selects the highest published thinking effort for the model and stores the v2 benchmark slice separately.
Multimodal3 benchmarks
CharXivProvider exact

CharXiv Reasoning

77.4%Weighted 25%
Source: Anthropic: Claude Sonnet 4.6 system cardProvenance: Anthropic reports CharXiv at 77.4% with the image-cropping tool in the Claude Sonnet 4.6 system card.
AA-MMMU-ProReported

Artificial Analysis MMMU-Pro

70.6%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Design Arena WebsiteReported

Design Arena Website Elo

1317Display only
Source: OpenRouter model benchmarksProvenance: Display-only Design Arena Website Elo synced from OpenRouter model benchmark metadata. It is excluded from BenchLM weighted scoring.
Inst. Following1 benchmark
AA-IFBenchReported

Artificial Analysis IFBench

41.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.

Frequently Asked Questions

How does Claude Sonnet 4.6 perform overall in AI benchmarks?

Claude Sonnet 4.6 currently ranks #32 out of 200 models on BenchLM's provisional leaderboard with an overall score of 65.07. It also ranks #27 out of 99 on the verified leaderboard. It is created by Anthropic. Its published context window is 200K.

Is Claude Sonnet 4.6 good for knowledge and understanding?

Claude Sonnet 4.6 ranks #15 out of 52 models in knowledge and understanding benchmarks with an average score of 79.7. There are stronger options in this category.

Is Claude Sonnet 4.6 good for coding and programming?

Claude Sonnet 4.6 ranks #23 out of 122 models in coding and programming benchmarks with an average score of 59.6. There are stronger options in this category.

Is Claude Sonnet 4.6 good for mathematics?

Claude Sonnet 4.6 has visible benchmark coverage in mathematics, but BenchLM does not currently assign it a global category rank there.

Is Claude Sonnet 4.6 good for reasoning and logic?

Claude Sonnet 4.6 has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.

Is Claude Sonnet 4.6 good for agentic tool use and computer tasks?

Claude Sonnet 4.6 ranks #42 out of 119 models in agentic tool use and computer tasks benchmarks with an average score of 51.2. There are stronger options in this category.

Is Claude Sonnet 4.6 good for multimodal and grounded tasks?

Claude Sonnet 4.6 has visible benchmark coverage in multimodal and grounded tasks, but BenchLM does not currently assign it a global category rank there.

Is Claude Sonnet 4.6 good for instruction following?

Claude Sonnet 4.6 has visible benchmark coverage in instruction following, but BenchLM does not currently assign it a global category rank there.

Does Claude Sonnet 4.6 have full benchmark coverage on BenchLM?

Not yet. Claude Sonnet 4.6 currently has 33 published benchmark scores out of the 321 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of Claude Sonnet 4.6?

Claude Sonnet 4.6 has a published context window of 200K, which determines how much text it can process in a single interaction.

Last updated: July 20, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Choose with this week’s evidence

Join 2,000+ readers for ranking moves, new releases, pricing changes, and the evidence behind them.

Free. One email per week.