Enter opens the highlighted result. Esc closes.

Claude Sonnet 4.6 vs Claude Sonnet 5

Claude Sonnet 5 leads Claude Sonnet 4.6 on coding, 57.7 to 57.3 on the Vibe Coding Index, and Claude Sonnet 5 costs 1.5x less per output token ($10.00 vs $15.00 per million). Updated October 22, 2026; 104 benchmarks have a result for both.

Replace Claude Sonnet 4.6 Replace Claude Sonnet 5
Claude Sonnet 4.6Anthropic
Coding
57.3
Overall
60.3
Output $/M
$15.00
42
Claude Sonnet 5Anthropic
Coding
57.7
Overall
63.2
Output $/M
$10.00
61
vs boards won of 104 shared, 1 tied

Claude Sonnet 4.6 or Claude Sonnet 5: which to pick

Claude Sonnet 4.6 and Claude Sonnet 5 are level for coding, 57.3 against 57.7 on the Vibe Coding Index, so price and context decide.

Pick Claude Sonnet 4.6 for a lower cost per solved task ($1.68 against $1.87) and work measured by ProgramBench (Vals AI) or SkillsBench (Vals AI), where it leads. Pick Claude Sonnet 5 for output tokens 1.5x cheaper ($10.00 vs $15.00 per million) and work measured by LiveBench coding or LMArena WebDev, where it leads.

Model AClaude Sonnet 4.6Anthropic57.3/1005 independent benchmarks$1.68 per solved taskModel BClaude Sonnet 5Anthropic57.7/1006 independent benchmarks$1.87 per solved task
Shared results
104
Claude Sonnet 4.6 only
19
Claude Sonnet 5 only
19
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Claude Sonnet 4.6 57.3 from 5 benchmarks, Claude Sonnet 5 57.7 from 6 benchmarks.
Widest category
Math, 10.5 points apart on average over 4 shared benchmarks.
Widest single gap
ProofBench: Claude Sonnet 4.6 45.0% against Claude Sonnet 5 77.0%, 32.0 points apart on a 0 to 100 scale.
Output price
Claude Sonnet 4.6 $15.00 against Claude Sonnet 5 $10.00 per million output tokens, 1.5x apart.
Context window
Both take 1M tokens.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryClaude Sonnet 4.6BarsClaude Sonnet 5Ties
Coding414
Intelligence02
Agentic work214
Reasoning67
Math13
Knowledge64
Web design441 tie
General1913
All shared boards42611 tie
CodingLike-for-likeSix of these boards feed the Vibe Coding Index414
Coding benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
LiveBench coding66.6%74.7%Claude Sonnet 5
LMArena WebDev15231540Claude Sonnet 5
AA Coding Index63.071.5Claude Sonnet 5
LiveCodeBench (Vals AI)82.1%82.4%Claude Sonnet 5
SWE-bench Verified (Vals AI)77.4%79.6%Claude Sonnet 5
Terminal-Bench 4 (Vals AI)Not measured10.6%Only Claude Sonnet 5 measured
Vibe Code Bench (Vals AI)55.8%81.3%Claude Sonnet 5
Code migration (Vals AI)39.9%44.4%Claude Sonnet 5
ProgramBench (Vals AI)0.5%0.0%Claude Sonnet 4.6
IOI (Vals AI)Not measured45.0%Only Claude Sonnet 5 measured
Terminal-Bench 2.1 (Vals AI)57.3%74.5%Claude Sonnet 5
Vibe Code Bench 1-100 (Vals AI)Not measured13.8%Only Claude Sonnet 5 measured
SkillsBench (Vals AI)49.0%46.5%Claude Sonnet 4.6
CyberBench (Vals AI)Not measured70.3%Only Claude Sonnet 5 measured
SciCode50.1%54.3%Claude Sonnet 5
Terminal-Bench Hard (Artificial Analysis run)53.0%Not measuredOnly Claude Sonnet 4.6 measured
Terminal-Bench53.4%Not measuredOnly Claude Sonnet 4.6 measured
FrontierCode24.3%42.7%Claude Sonnet 5
GSO-BenchNot measured37.3%Only Claude Sonnet 5 measured
WeirdML66.1%68.8%Claude Sonnet 5
ALE-BenchNot measured1463.1Only Claude Sonnet 5 measured
LMArena Text: coding15031483Claude Sonnet 4.6
LMArena Text: software and IT14911474Claude Sonnet 4.6
LMArena WebDev: React15121541Claude Sonnet 5
LMArena WebDev: HTML15271527Claude Sonnet 5
LMArena WebDev: from an image15291540Claude Sonnet 5
SciCodeNot measured10.7%Only Claude Sonnet 5 measured
SWE-Bench Pro V2Not measured93.2%Only Claude Sonnet 5 measured
SWE Atlas: Codebase QnA31.2%Not measuredOnly Claude Sonnet 4.6 measured
SWE Atlas: Refactoring32.2%Not measuredOnly Claude Sonnet 4.6 measured
SWE Atlas: Test Writing31.8%Not measuredOnly Claude Sonnet 4.6 measured
Neither model has a result on the other 12 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
AA Intelligence Index30.138.2Claude Sonnet 5
Epoch Capabilities Index152.8156.9Claude Sonnet 5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index214
Agentic work benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
AA Agentic Index31.843.6Claude Sonnet 5
Design Arena: agents building web apps11901245Claude Sonnet 5
Design Arena: agents building full-stack apps12131249Claude Sonnet 5
Design Arena: agents building mobile apps12111215Claude Sonnet 5
Design Arena: agents building Android apps11961226Claude Sonnet 5
Design Arena: agents building Godot games12261225Claude Sonnet 4.6
Design Arena: agents building games11771227Claude Sonnet 5
Design Arena: HTML slidesNot measured1157Only Claude Sonnet 5 measured
Design Arena: PowerPoint slidesNot measured1190Only Claude Sonnet 5 measured
Design Arena: spreadsheetsNot measured1079Only Claude Sonnet 5 measured
LiveBench research63.6%82.8%Claude Sonnet 5
LiveBench browser68.0%72.0%Claude Sonnet 5
LiveBench long agentic23.1%39.0%Claude Sonnet 5
SREBench (Vals AI)Not measured0.4%Only Claude Sonnet 5 measured
Finance Agent (Vals AI)51.0%53.9%Claude Sonnet 5
Tax Agent Bench (Vals AI)25.8%29.4%Claude Sonnet 5
Legal Research Bench (Vals AI)38.5%41.8%Claude Sonnet 5
Excel Modeling Benchmark (Vals AI)60.2%66.3%Claude Sonnet 5
Terminal-Bench-Science (Vals AI)Not measured5.7%Only Claude Sonnet 5 measured
τ²-Bench Telecom79.5%Not measuredOnly Claude Sonnet 4.6 measured
APEX-Agents43.0%54.5%Claude Sonnet 5
OSWorld 2.09.3%Not measuredOnly Claude Sonnet 4.6 measured
DeepResearch Bench54.9%Not measuredOnly Claude Sonnet 4.6 measured
OSWorld72.1%Not measuredOnly Claude Sonnet 4.6 measured
Vending-Bench 2$7,204$6,378Claude Sonnet 4.6
MCP Atlas69.5%Not measuredOnly Claude Sonnet 4.6 measured
Neither model has a result on the other 16 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index67
Reasoning benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
GPQA Diamond85.6%88.9%Claude Sonnet 5
LiveBench reasoning86.4%88.7%Claude Sonnet 5
MMMU Pro (Vals AI)83.6%83.0%Claude Sonnet 4.6
GPQA Diamond (Artificial Analysis run)87.5%91.1%Claude Sonnet 5
Humanity's Last Exam (Artificial Analysis run)33.6%41.3%Claude Sonnet 5
GPQA Diamond (Epoch AI run)87.4%90.5%Claude Sonnet 5
ARC-AGI86.5%Not measuredOnly Claude Sonnet 4.6 measured
ARC-AGI-260.4%Not measuredOnly Claude Sonnet 4.6 measured
SimpleBenchNot measured60.6%Only Claude Sonnet 5 measured
Chess puzzles23.0%45.0%Claude Sonnet 5
Mystery game puzzles16.0%35.0%Claude Sonnet 5
LMArena Text: hard prompts14851461Claude Sonnet 4.6
LMArena Text: expert15001489Claude Sonnet 4.6
LMArena Text: hard prompts in English14931464Claude Sonnet 4.6
LMArena Vision: diagrams13061293Claude Sonnet 4.6
LMArena Vision: homework13111311Claude Sonnet 4.6
Neither model has a result on the other 7 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index13
Math benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
AIME92.3%Not measuredOnly Claude Sonnet 4.6 measured
OTIS Mock AIME85.8%94.7%Claude Sonnet 5
FrontierMath, tiers 1 to 3Not measured65.6%Only Claude Sonnet 5 measured
FrontierMath, tier 4Not measured29.3%Only Claude Sonnet 5 measured
ProofBench45.0%77.0%Claude Sonnet 5
LMArena Text: math14621467Claude Sonnet 5
LMArena Text: mathematical work14831481Claude Sonnet 4.6
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index64
Knowledge benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
MMLU Pro87.3%87.5%Claude Sonnet 5
LegalBench (Vals AI)82.1%83.9%Claude Sonnet 5
CaseLaw (Vals AI)64.0%Not measuredOnly Claude Sonnet 4.6 measured
CorpFin (Vals AI)65.3%67.0%Claude Sonnet 5
TaxEval (Vals AI)77.1%75.6%Claude Sonnet 4.6
MortgageTax (Vals AI)67.7%70.0%Claude Sonnet 5
MedQA (Vals AI)92.1%Not measuredOnly Claude Sonnet 4.6 measured
MedCode (Vals AI)Not measured47.5%Only Claude Sonnet 5 measured
SimpleQA Verified35.5%33.7%Claude Sonnet 4.6
LMArena Text: science14741460Claude Sonnet 4.6
LMArena Text: medicine and healthcare14511442Claude Sonnet 4.6
LMArena Text: legal and government14701454Claude Sonnet 4.6
LMArena Text: business and finance14641436Claude Sonnet 4.6
FACTS Benchmark Suite37.0%Not measuredOnly Claude Sonnet 4.6 measured
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index44
Web design benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
Design Arena: websites12841274Claude Sonnet 4.6
Design Arena: UI components12691277Claude Sonnet 5
Design Arena: game dev12581276Claude Sonnet 5
Design Arena: data viz12781252Claude Sonnet 4.6
Design Arena: 3D12341261Claude Sonnet 5
Design Arena: all code categories12751275Tie
Design Arena: SVG11961191Claude Sonnet 4.6
Design Arena: ASCII art12351206Claude Sonnet 4.6
LiveBench design71.8%82.0%Claude Sonnet 5
GeneralLike-for-likePublic boards, shown as published, kept out of the index1913
General benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
LiveBench data analysis76.1%71.7%Claude Sonnet 4.6
Public Benefits Bench (Vals AI)62.5%66.0%Claude Sonnet 5
SAGE (Vals AI)46.6%48.9%Claude Sonnet 5
MedScribe (Vals AI)Not measured76.0%Only Claude Sonnet 5 measured
Vals Multimodal Index60.6%68.8%Claude Sonnet 5
IFBench56.6%Not measuredOnly Claude Sonnet 4.6 measured
Long context reasoning (AA-LCR)80.0%82.0%Claude Sonnet 5
DTBench89.9%92.5%Claude Sonnet 5
LMCA46.5%50.0%Claude Sonnet 5
Surface Evolver BenchNot measured60.0%Only Claude Sonnet 5 measured
ExploitBench23.6%Not measuredOnly Claude Sonnet 4.6 measured
LMArena Text14581442Claude Sonnet 4.6
LMArena Text: instruction following14761452Claude Sonnet 4.6
LMArena Text: creative writing14351416Claude Sonnet 4.6
LMArena Vision12831273Claude Sonnet 4.6
LMArena Text: English14721450Claude Sonnet 4.6
LMArena Text: non-English14401431Claude Sonnet 4.6
LMArena Text: Chinese14911478Claude Sonnet 4.6
LMArena Text: German14281442Claude Sonnet 5
LMArena Text: Japanese14201422Claude Sonnet 5
LMArena Text: Korean14101412Claude Sonnet 5
LMArena Text: French14651463Claude Sonnet 4.6
LMArena Text: Spanish14641438Claude Sonnet 4.6
LMArena Text: Russian14401450Claude Sonnet 5
LMArena Text: Polish14431425Claude Sonnet 4.6
LMArena Text: multi-turn14641453Claude Sonnet 4.6
LMArena Text: longer queries14791462Claude Sonnet 4.6
LMArena Text: writing and language14441436Claude Sonnet 4.6
LMArena Text: entertainment and media14311411Claude Sonnet 4.6
LMArena Vision: English12981275Claude Sonnet 4.6
LMArena Vision: Chinese13071331Claude Sonnet 5
LMArena Vision: reading text12961288Claude Sonnet 4.6
LMArena Vision: humor12481261Claude Sonnet 5
LMArena Vision: recognizing things1195Not measuredOnly Claude Sonnet 4.6 measured
LMArena Vision: creative writing12721246Claude Sonnet 4.6
LMArena Document14821466Claude Sonnet 4.6
EQ-Bench 412071236Claude Sonnet 5
Neither model has a result on the other 11 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Sonnet 4.6 and Claude Sonnet 5
BenchmarkClaude Sonnet 4.6Claude Sonnet 5Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Claude Sonnet 4.6$0.011Fits in one request
Claude Sonnet 5$0.0070Fits in one request

Claude Sonnet 5 costs less

Repository review

50K input + 3K output tokens

Claude Sonnet 4.6$0.195Fits in one request
Claude Sonnet 5$0.130Fits in one request

Claude Sonnet 5 costs less

Agent loop

200K input + 10K output tokens

Claude Sonnet 4.6$0.750Fits in one request
Claude Sonnet 5$0.500Fits in one request

Claude Sonnet 5 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Claude Sonnet 4.6 and Claude Sonnet 5
SpecificationAnthropic Claude Sonnet 4.6Anthropic Claude Sonnet 5
MakerAnthropicAnthropic
WeightsNo public weights listedNo public weights listed
ReleasedFebruary 17, 2026June 30, 2026
Context window1M1M
Input $/Mtok$3.00$2.00
Output $/Mtok$15.00$10.00
Coding index57.357.7
Independent measurements behind it56
Cost per solved task$1.68$1.87
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Sonnet 4.6 better than Claude Sonnet 5 for coding?

Claude Sonnet 4.6 and Claude Sonnet 5 are level for coding, 57.3 against 57.7 on the Vibe Coding Index, so price and context decide. Pick Claude Sonnet 4.6 for a lower cost per solved task ($1.68 against $1.87) and work measured by ProgramBench (Vals AI) or SkillsBench (Vals AI), where it leads. Pick Claude Sonnet 5 for output tokens 1.5x cheaper ($10.00 vs $15.00 per million) and work measured by LiveBench coding or LMArena WebDev, where it leads.

Which is cheaper, Claude Sonnet 4.6 or Claude Sonnet 5?

Claude Sonnet 4.6 costs $3.00 per million input tokens and $15.00 per million output tokens; Claude Sonnet 5 costs $2.00 and $10.00. Per solved SWE-bench Verified task, Claude Sonnet 4.6 costs $1.68 and Claude Sonnet 5 $1.87.

Which has the bigger context window, Claude Sonnet 4.6 or Claude Sonnet 5?

Neither: both take 1M tokens.

How many benchmarks is the Claude Sonnet 4.6 vs Claude Sonnet 5 comparison based on?

104 benchmarks have a result for both models as of October 22, 2026. Claude Sonnet 4.6's coding index rests on 5 independent benchmarks and Claude Sonnet 5's on 6. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/claude-sonnet-4-6-vs-claude-sonnet-5" title="Claude Sonnet 4.6 vs Claude Sonnet 5, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/claude-sonnet-4-6-vs-claude-sonnet-5">LLMBenchmark.ai, Claude Sonnet 4.6 vs Claude Sonnet 5</a></p>