Skip to main content
BenchLM

Data as of September 29, 2026 · How the score is built

Superseded.Anthropic has newer models in this line:Claude Sonnet 5.5
SupersededProprietaryReasoning

Released Jun 30, 20261M context

Claude Sonnet 5

Decision readingClaude Sonnet 5 scores 67 out of 100 and ranks #23 of 209. This profile shows 26 source-displayable benchmark rows; its strongest eligible category is Agentic at #12. API pricing is $2 input and $10 output per million tokens, with cached input at $0.2.

Released Jun 30, 2026 — see all recent releases

Find alternativesCompare Claude Sonnet 5
Share or export
Share on XLinkedIn
Social cardBenchmarks CSVJSON

Decision snapshot

Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.

Capability

67/100

field median 50.5#23 of 209 ranked models

Public

#23of 209

Verified #15 of 74

Price

$2input / $10 output

input median $1cached $0.20 · batch + cache $0.10 · blended $6

Speed

75tok/s

field median 93 tok/sFirst token 202.74 s

Context

1Mtokens

field median 256,000Maximum output length is tracked separately

Strongest published evidence

Agentic ranks #12. Particularly useful for coding agents, browser research, and computer-use workflows.

Validate before choosing

26 published rows leave some tracked benchmark slots empty. Knowledge is its lowest eligible category at #26.

Source-linked · 26 displayable benchmark rows

Follow model changes

Check the published record for the model and serving route you use.

Check this model

Category score record

Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #12 of 117Percentile 91stWeight 22%7 benchmarksVerified
64.6
CodingRank #19 of 142Percentile 87thWeight 20%11 benchmarksVerified
59.8
ReasoningWeight 17%0 benchmarksNot measured
Not measured
MultimodalRank #15 of 50Percentile 71stWeight 12%2 benchmarksVerified
78.4
KnowledgeRank #26 of 168Percentile 85thWeight 12%6 benchmarksVerified
64.5
MultilingualWeight 7%0 benchmarksNot measured
Not measured
Inst. FollowingWeight 5%0 benchmarksNot measured
Not measured
MathWeight 5%0 benchmarksNot measured
Not measured

26 of 486 tracked benchmark slots have displayable evidence · bars run 0–100

Coverage details

How much of this is verified

Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.

  1. Agentic7/7 verified
  2. Coding11/11 verified
  3. ReasoningNot measured
  4. Multimodal2/2 verified
  5. Knowledge6/6 verified
  6. MultilingualNot measured
  7. Inst. FollowingNot measured
  8. MathNot measured
Verified sourceProvisionalNot measured

Capability shape

Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.

Claude Sonnet 5 category percentile values

  • Agentic91st percentile
  • Coding87th percentile
  • ReasoningNot eligible
  • Multimodal71st percentile
  • Knowledge85th percentile
  • MultilingualNot eligible
  • Instruction followingNot eligible
  • MathNot eligible

The dashed outline is median of 6 nearest peers.

Eligible category ranks

  1. Agentic#12/117
  2. Coding#19/142
  3. ReasoningNot ranked
  4. Multimodal#15/50
  5. Knowledge#26/168
  6. MultilingualNot ranked
  7. Inst. FollowingNot ranked
  8. MathNot ranked
Top decileTop quartileMid-fieldNot eligible

Benchmark ledger

Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.

Coding11 rows
Coding benchmark values, best verified comparison, weight, and source status
SWE-bench ProScore63.2%Versus best verified row

Best verified: Claude Opus 5.5 · 89.9%

Gap26.7 behindWeight26% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
CursorBench 3.2Score61.5%Versus best verified row

Best verified: Claude Fable 5.1 · 73.4%

Gap11.9 behindWeight10% ref. weight
Benchmark exact
Cursor evals: CursorBench 3.2
LiveCodeBench (Vals)LiveCodeBench, Vals AI runScore82.4%Versus best verified row

Best verified: Claude Fable 5.1 · 90.5%

Gap8.1 behindWeight8% ref. weight
Verified
Vals AI: LiveCodeBench leaderboard
SWE MultilingualScore78.3%Versus best verified row

Best verified: Claude Opus 5.5 · 93.9%

Gap15.6 behindWeight5% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
FrontierCode 1.1 MainScore42.7%Versus best verified row

Best verified: Claude Opus 5.5 · 54.4%

Gap11.7 behindWeight4% ref. weight
Benchmark exact
Cognition: FrontierCode 1.1
Terminal-Bench 2.1Score80.4%Versus best verified row

Best verified: SWE-2 · 92.8%

Gap12.4 behindWeightScored in Agentic
Provider exact
Anthropic: Claude Sonnet 5 system card
SWE-bench VerifiedSoftware Engineering Benchmark VerifiedScore85.2%Versus best verified row

Best verified: Claude Opus 5 · 96%

Gap10.8 behindWeightDisplay only
Provider exact
Anthropic: Claude Sonnet 5 system card
SWE MultimodalSWE-bench MultimodalScore28.1%Versus best verified row

Best verified: Claude Opus 5.5 · 61.4%

Gap33.3 behindWeightDisplay only
Provider exact
Anthropic: Claude Sonnet 5 system card
VulcanBench CII v1VulcanBench Coding Intelligence Index v1Score89.2%Versus best verified row

Best verified: Claude Opus 5 · 96.4%

Gap7.2 behindWeightDisplay only
Benchmark exact
VulcanBench CII v1 frontier results
SWE-bench (Vals)SWE-bench, Vals AI runScore79.6%Versus best verified row

Best verified: Claude Opus 5 · 97.0%

Gap17.4 behindWeightDisplay only
Verified
Vals AI: SWE-bench leaderboard
CursorBench 4.0Score34.1%Versus best verified row

Best verified: Claude Opus 5.5 · 57.8%

Gap23.7 behindWeightDisplay only
Benchmark exact
Cursor evals: CursorBench 4.0
Agentic7 rows
Agentic benchmark values, best verified comparison, weight, and source status
Terminal-Bench 2.1Score80.4%Versus best verified row

Best verified: SWE-2 · 92.8%

Gap12.4 behindWeight8% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
BrowseCompScore84.7%Versus best verified row

Best verified: Atria Dawn Preview · 92.5%

Gap7.8 behindWeight8% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
OSWorld-VerifiedScore81.2%Versus best verified row

Best verified: Qwen3.8 Max · 86.1%

Gap4.9 behindWeight6% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
Terminal-Bench 2.1 (Vals)Terminal-Bench 2.1, Vals AI runScore74.5%Versus best verified row

Best verified: GPT-6 Astra · 87.3%

Gap12.8 behindWeight3% ref. weight
Verified
Vals AI: Terminal-Bench 2.1 leaderboard
Terminal-Bench 3.0Score14.6%Versus best verified row

Best verified: Claude Opus 5 · 42.7%

Gap28.1 behindWeight3% ref. weight
Benchmark exact
Terminal-Bench 3.0 leaderboard
HLE w/ toolsHumanity's Last Exam with toolsScore57.4%Versus best verified row

Best verified: Claude Opus 5.5 · 67.7%

Gap10.3 behindWeight3% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
ApprenticeBenchApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable jobScore16%Versus best verified row

Best verified: Claude Fable 5.1 · 72%

Gap56 behindWeightDisplay only
Benchmark exact
NeoCognition: ApprenticeBench GUI results
Multimodal2 rows
Multimodal benchmark values, best verified comparison, weight, and source status
CharXivCharXiv ReasoningScore88.3%Versus best verified row

Best verified: Qwen3.8 Max · 93.5%

Gap5.2 behindWeightWeighted 20%
Provider exact
Anthropic: Claude Sonnet 5 system card
CharXiv w/o toolsCharXiv Reasoning without toolsScore77%Versus best verified row

Best verified: Claude Mythos 5 · 88.9%

Gap11.9 behindWeightDisplay only
Provider exact
Anthropic: Claude Sonnet 5 system card
Knowledge6 rows
Knowledge benchmark values, best verified comparison, weight, and source status
HLEHumanity's Last ExamScore57.4%Versus best verified row

Best verified: Claude Fable 5.1 · 65%

Gap7.6 behindWeight44% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
HLE w/o toolsHumanity's Last Exam without toolsScore43.2%Versus best verified row

Best verified: Claude Opus 5.5 · 64.4%

Gap21.2 behindWeight7% ref. weight
Provider exact
Anthropic: Claude Sonnet 5 system card
MMLU-Pro (Vals)MMLU-Pro, Vals AI runScore87.5%Versus best verified row

Best verified: Claude Fable 5.1 · 92.4%

Gap4.9 behindWeight6% ref. weight
Verified
Vals AI: MMLU Pro leaderboard
GPQA Diamond (Vals)GPQA Diamond, Vals AI runScore88.9%Versus best verified row

Best verified: Gemini 3.1 Pro · 95.5%

Gap6.6 behindWeight2% ref. weight
Verified
Vals AI: GPQA Diamond leaderboard
HLE-VerifiedScore31.0%Versus best verified row

Best verified: Gemini 3.8 Flash · 54.9%

Gap23.9 behindWeightDisplay only
Provider exact
Google DeepMind: Gemini 3.8 Flash model card
LABBench2LABBench2: An Improved Benchmark for AI Systems Performing Biology ResearchScore80.1%Versus best verified row

Best verified: Gemini 3.8 Flash · 86.2%

Gap6.1 behindWeightDisplay only
Provider exact
Google DeepMind: Gemini 3.8 Flash model card

Bars run 0–100; the dark tick marks the best source-verified value

All 26 rows

What it costs to get this score

Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.

Current modelExplore all models

Claude Sonnet 5 · 67.0 score · $6 blended per million tokens

The chart opens on the current model. Scroll horizontally to inspect the full price axis.

2030405060708090100$0.10$0.50$1$5$10$25$50$100↘ frontierClaude Sonnet 5

Horizontal: blended price per million tokens, log scale · Vertical: public score

Price vs performance

Lineage

The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.

  1. February 2026

    Claude Sonnet 4.6

    Score 56.3 · $3 / $15

  2. Jun 30, 2026 · you are here

    Claude Sonnet 5

    Score 67.0 · $2 / $10

  3. Sep 28, 2026

    Claude Sonnet 5.5

    Score 80.5 · $2 / $10

Base entry

Radar

Claude Sonnet 5 release history

Full release history

Radar confirmed these at the source. Use Claude Sonnet 5 in your work? Explore Radar to follow supported changes and choose your alerts.

Radar

Spec sheet

Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.

API model ID
claude-sonnet-5Anthropic model overview
Context window
1MAnthropic model overview
Maximum output
Not sourced yet
Knowledge cutoff
Not sourced yet
Input modalities
text, imageAnthropic model overview
Output modalities
textAnthropic model overview
Parameters
Not disclosed by the provider
Availability
Claude APIAnthropic model overview
Cloud regions
Not tracked yet
Lifecycle
activeAnthropic model overview
API capabilities
Tool calling, structured outputs, and batch support are not tracked yet
Prompt caching
Published at $0.20 per million cached input tokensClaude API pricing
Self-host
Weights are not published
Rate limits
Not tracked yet

How to read this profile

The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.

Claude Sonnet 5 ranks #23 of 209 on the public leaderboard with a score of 66.98/100. Its source-verified position is #15 of 74.

Claude Sonnet 5 is a proprietary model with a 1M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.

Official exact-value snapshot from Anthropic's June 30, 2026 Claude Sonnet 5 launch post and system card, with HLE-Verified and LABBench2 comparison rows from Google's Gemini 3.8 Flash model card. We map exact rows that align with existing keys and keep the two Google comparison results display-only. Cost-performance curves and safety-only measurements remain out of scored fields.

Its explicit predecessor is Claude Sonnet 4.6. 26 of 486 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.

Its strongest eligible category is Agentic at #12, while its lowest eligible position is Knowledge at #26. particularly useful for coding agents, browser research, and computer-use workflows.

Last updated September 29, 2026. Runtime fields remain blank until a sourced snapshot exists.

Questions

How does Claude Sonnet 5 perform overall in AI benchmarks?

Claude Sonnet 5 ranks #23 out of 209 models on the public BenchAlign leaderboard, with a score of 66.98/100. Its evidence status is Supported, and this profile shows 26 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.

Is Claude Sonnet 5 good for knowledge and understanding?

Claude Sonnet 5 ranks #26 out of 168 eligible models for knowledge and understanding, with a public category score of 64.5/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Claude Sonnet 5 good for coding and programming?

Claude Sonnet 5 ranks #19 out of 142 eligible models for coding and programming, with a public category score of 59.8/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Claude Sonnet 5 good for agentic tool use and computer tasks?

Claude Sonnet 5 ranks #12 out of 117 eligible models for agentic tool use and computer tasks, with a public category score of 64.6/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Claude Sonnet 5 good for multimodal and grounded tasks?

Claude Sonnet 5 ranks #15 out of 50 eligible models for multimodal and grounded tasks, with a public category score of 78.4/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Does Claude Sonnet 5 have full benchmark coverage on BenchLM?

No. Claude Sonnet 5 currently has 43 source-displayable rows across 486 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.

What is the context window size of Claude Sonnet 5?

Claude Sonnet 5 has a documented context window of 1M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.

Watch Claude Sonnet 5 in the weekly brief

Get one weekly email when material rank, price, availability, or benchmark evidence changes are worth revisiting.

Read a sample issue

Join 2,000+ readers.

Compare Claude Sonnet 5 with every tracked model512 comparisons