Enter opens the highlighted result. Esc closes.

Claude Sonnet 5 vs GPT-5.4

GPT-5.4 leads Claude Sonnet 5 on coding, 64.4 to 57.7 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 21, 2026; 91 benchmarks have a result for both.

Replace Claude Sonnet 5 Replace GPT-5.4
Claude Sonnet 5Anthropic
Coding
57.7
Overall
63.2
Output $/M
$10.00
31
GPT-5.4OpenAI
Coding
64.4
Overall
69.6
Output $/M
$15.00
59
vs boards won of 91 shared, 1 tied

Claude Sonnet 5 or GPT-5.4: which to pick

Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 57.7 for Claude Sonnet 5, though on 3 benchmarks against 6.

Pick Claude Sonnet 5 for output tokens 1.5x cheaper ($10.00 vs $15.00 per million) and work measured by LMArena WebDev or AA Coding Index, where it leads.

Model AClaude Sonnet 5Anthropic57.7/1006 independent benchmarks$1.87 per solved taskModel BGPT-5.4OpenAI64.4/1003 independent benchmarks$1.02 per solved task
Shared results
91
Claude Sonnet 5 only
30
GPT-5.4 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 5 57.7 from 6 benchmarks, GPT-5.4 64.4 from 3 benchmarks.
Widest category
Agentic work, 15.3 points apart on average over 7 shared benchmarks.
Widest single gap
Design Arena: agents building full-stack apps: Claude Sonnet 5 1249 against GPT-5.4 1017, 29.0 points apart on a 0 to 100 scale.
Output price
Claude Sonnet 5 $10.00 against GPT-5.4 $15.00 per million output tokens, 1.5x apart.
Context window
Claude Sonnet 5 takes 1M tokens, GPT-5.4 1.05M.

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 5BarsGPT-5.4Ties
Coding97
Intelligence02
Agentic work70
Reasoning012
Math15
Knowledge56
Web design80
General1271 tie
All shared boards31591 tie
CodingLike-for-likeSix of these boards feed the Vibe Coding Index97
Coding benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
LiveBench coding74.7%Not measuredOnly Claude Sonnet 5 measured
LMArena WebDev15401464Claude Sonnet 5
AA Coding Index71.571.1Claude Sonnet 5
LiveCodeBench (Vals AI)82.4%84.1%GPT-5.4
SWE-bench Verified (Vals AI)79.6%78.2%Claude Sonnet 5
Terminal-Bench 4 (Vals AI)10.6%Not measuredOnly Claude Sonnet 5 measured
Vibe Code Bench (Vals AI)81.3%67.4%Claude Sonnet 5
Code migration (Vals AI)44.4%35.0%Claude Sonnet 5
ProgramBench (Vals AI)0.0%0.5%GPT-5.4
IOI (Vals AI)45.0%Not measuredOnly Claude Sonnet 5 measured
Terminal-Bench 2.1 (Vals AI)74.5%Not measuredOnly Claude Sonnet 5 measured
Vibe Code Bench 1-100 (Vals AI)13.8%Not measuredOnly Claude Sonnet 5 measured
SkillsBench (Vals AI)46.5%51.7%GPT-5.4
CyberBench (Vals AI)70.3%Not measuredOnly Claude Sonnet 5 measured
SciCode54.3%Not measuredOnly Claude Sonnet 5 measured
Terminal-Bench Hard (Artificial Analysis run)Not measured57.6%Only GPT-5.4 measured
Terminal-BenchNot measured81.8%Only GPT-5.4 measured
FrontierCode42.7%Not measuredOnly Claude Sonnet 5 measured
GSO-Bench37.3%31.4%Claude Sonnet 5
WeirdML68.8%77.7%GPT-5.4
MirrorCodeNot measured15.6%Only GPT-5.4 measured
ALE-Bench1463.11607.0GPT-5.4
LMArena Text: coding14831496GPT-5.4
LMArena Text: software and IT14741491GPT-5.4
LMArena WebDev: React15411454Claude Sonnet 5
LMArena WebDev: HTML15271478Claude Sonnet 5
LMArena WebDev: from an image15401436Claude Sonnet 5
SciCode10.7%Not measuredOnly Claude Sonnet 5 measured
Neither model has a result on the other 11 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
AA Intelligence Index38.239.0GPT-5.4
Epoch Capabilities Index156.9157.3GPT-5.4
Agentic workLike-for-likePublic boards, shown as published, kept out of the index70
Agentic work benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
AA Agentic Index43.6Not measuredOnly Claude Sonnet 5 measured
Design Arena: agents building web apps12451059Claude Sonnet 5
Design Arena: agents building full-stack apps12491017Claude Sonnet 5
Design Arena: agents building mobile apps12151087Claude Sonnet 5
Design Arena: agents building Android apps12251030Claude Sonnet 5
Design Arena: agents building Godot games12251135Claude Sonnet 5
Design Arena: agents building games1227Not measuredOnly Claude Sonnet 5 measured
Design Arena: HTML slides1218Not measuredOnly Claude Sonnet 5 measured
Design Arena: PowerPoint slides1216Not measuredOnly Claude Sonnet 5 measured
Design Arena: spreadsheets1021Not measuredOnly Claude Sonnet 5 measured
LiveBench research82.8%Not measuredOnly Claude Sonnet 5 measured
LiveBench browser72.0%Not measuredOnly Claude Sonnet 5 measured
LiveBench long agentic39.0%Not measuredOnly Claude Sonnet 5 measured
SREBench (Vals AI)0.4%Not measuredOnly Claude Sonnet 5 measured
Finance Agent (Vals AI)53.9%Not measuredOnly Claude Sonnet 5 measured
Tax Agent Bench (Vals AI)29.4%Not measuredOnly Claude Sonnet 5 measured
Legal Research Bench (Vals AI)41.8%Not measuredOnly Claude Sonnet 5 measured
Excel Modeling Benchmark (Vals AI)66.3%Not measuredOnly Claude Sonnet 5 measured
Terminal-Bench-Science (Vals AI)5.7%Not measuredOnly Claude Sonnet 5 measured
τ²-Bench TelecomNot measured87.1%Only GPT-5.4 measured
APEX-Agents54.5%52.4%Claude Sonnet 5
METR time horizonsNot measured74.3%Only GPT-5.4 measured
DeepResearch BenchNot measured35.1%Only GPT-5.4 measured
PostTrainBenchNot measured19.0%Only GPT-5.4 measured
Vending-Bench 2$6,378$6,144Claude Sonnet 5
Neither model has a result on the other 16 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index012
Reasoning benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
GPQA Diamond88.9%91.7%GPT-5.4
LiveBench reasoning88.7%Not measuredOnly Claude Sonnet 5 measured
MMMU Pro (Vals AI)83.0%87.5%GPT-5.4
GPQA Diamond (Artificial Analysis run)91.1%92.0%GPT-5.4
Humanity's Last Exam (Artificial Analysis run)41.3%43.7%GPT-5.4
GPQA Diamond (Epoch AI run)90.5%93.3%GPT-5.4
ARC-AGINot measured93.7%Only GPT-5.4 measured
ARC-AGI-2Not measured74.0%Only GPT-5.4 measured
SimpleBench60.6%Not measuredOnly Claude Sonnet 5 measured
Chess puzzles45.0%48.0%GPT-5.4
Mystery game puzzles35.0%37.0%GPT-5.4
EBR-benchNot measured25.4%Only GPT-5.4 measured
LMArena Text: hard prompts14611485GPT-5.4
LMArena Text: expert14891507GPT-5.4
LMArena Text: hard prompts in English14641479GPT-5.4
LMArena Vision: diagrams12931327GPT-5.4
LMArena Vision: homework13111336GPT-5.4
Kagi LLM BenchmarkNot measured63.8%Only GPT-5.4 measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index15
Math benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
AIMENot measured96.7%Only GPT-5.4 measured
OTIS Mock AIME94.7%97.8%GPT-5.4
FrontierMath, tiers 1 to 365.6%78.6%GPT-5.4
FrontierMath, tier 429.3%49.0%GPT-5.4
ProofBench77.0%56.0%Claude Sonnet 5
LMArena Text: math14671488GPT-5.4
LMArena Text: mathematical work14811497GPT-5.4
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index56
Knowledge benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
MMLU Pro87.5%87.5%Claude Sonnet 5
LegalBench (Vals AI)83.9%86.0%GPT-5.4
CaseLaw (Vals AI)Not measured63.8%Only GPT-5.4 measured
CorpFin (Vals AI)67.0%65.3%Claude Sonnet 5
TaxEval (Vals AI)75.6%74.0%Claude Sonnet 5
MortgageTax (Vals AI)70.0%68.3%Claude Sonnet 5
MedQA (Vals AI)Not measured96.1%Only GPT-5.4 measured
MedCode (Vals AI)47.5%41.3%Claude Sonnet 5
SimpleQA Verified33.7%45.1%GPT-5.4
LMArena Text: science14601479GPT-5.4
LMArena Text: medicine and healthcare14421458GPT-5.4
LMArena Text: legal and government14541486GPT-5.4
LMArena Text: business and finance14361474GPT-5.4
Neither model has a result on the other 5 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
Design Arena: websites12741199Claude Sonnet 5
Design Arena: UI components12801215Claude Sonnet 5
Design Arena: game dev12781195Claude Sonnet 5
Design Arena: data viz12511196Claude Sonnet 5
Design Arena: 3D12611066Claude Sonnet 5
Design Arena: all code categories12761180Claude Sonnet 5
Design Arena: SVG11931142Claude Sonnet 5
Design Arena: ASCII art12071188Claude Sonnet 5
LiveBench design82.0%Not measuredOnly Claude Sonnet 5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index127
General benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
LiveBench data analysis71.7%Not measuredOnly Claude Sonnet 5 measured
Public Benefits Bench (Vals AI)66.0%Not measuredOnly Claude Sonnet 5 measured
SAGE (Vals AI)48.9%43.3%Claude Sonnet 5
MedScribe (Vals AI)76.0%77.5%GPT-5.4
Vals Multimodal Index68.8%Not measuredOnly Claude Sonnet 5 measured
IFBenchNot measured73.9%Only GPT-5.4 measured
Long context reasoning (AA-LCR)82.0%82.0%Tie
DTBench92.5%94.4%GPT-5.4
LMCA50.0%52.0%GPT-5.4
CL-benchNot measured27.9%Only GPT-5.4 measured
Surface Evolver Bench60.0%Not measuredOnly Claude Sonnet 5 measured
Furniture assemblyNot measured37.5%Only GPT-5.4 measured
CL-bench LifeNot measured21.7%Only GPT-5.4 measured
LMArena Text14421469GPT-5.4
LMArena Text: instruction following14521469GPT-5.4
LMArena Text: creative writing14161439GPT-5.4
LMArena Vision12731303GPT-5.4
LMArena Text: English14501467GPT-5.4
LMArena Text: non-English14311465GPT-5.4
LMArena Text: Chinese14781520GPT-5.4
LMArena Text: German14421472GPT-5.4
LMArena Text: Japanese14221485GPT-5.4
LMArena Text: Korean14121449GPT-5.4
LMArena Text: French14631493GPT-5.4
LMArena Text: Spanish14381452GPT-5.4
LMArena Text: Russian14501480GPT-5.4
LMArena Text: Polish14251472GPT-5.4
LMArena Text: multi-turn14531482GPT-5.4
LMArena Text: longer queries14621473GPT-5.4
LMArena Text: writing and language14361458GPT-5.4
LMArena Text: entertainment and media14111438GPT-5.4
LMArena Vision: English12751301GPT-5.4
LMArena Vision: Chinese13311354GPT-5.4
LMArena Vision: reading text12881312GPT-5.4
LMArena Vision: humor12611284GPT-5.4
LMArena Vision: recognizing thingsNot measured1251Only GPT-5.4 measured
LMArena Vision: creative writing12461283GPT-5.4
LMArena Document14661471GPT-5.4
Neither model has a result on the other 8 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Sonnet 5 and GPT-5.4
BenchmarkClaude Sonnet 5GPT-5.4Leader
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 5$0.0070Fits in one request
GPT-5.4$0.010Fits in one request

Claude Sonnet 5 costs less

Repository review

50K input + 3K output tokens

Claude Sonnet 5$0.130Fits in one request
GPT-5.4$0.170Fits in one request

Claude Sonnet 5 costs less

Agent loop

200K input + 10K output tokens

Claude Sonnet 5$0.500Fits in one request
GPT-5.4$0.650Fits 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 5 and GPT-5.4
SpecificationAnthropic Claude Sonnet 5OpenAI GPT-5.4
MakerAnthropicOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedJune 30, 2026March 5, 2026
Context window1M1.05M
Input $/Mtok$2.00$2.50
Output $/Mtok$10.00$15.00
Coding index57.764.4
Independent measurements behind it63
Cost per solved task$1.87$1.02
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Sonnet 5 better than GPT-5.4 for coding?

Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 57.7 for Claude Sonnet 5, though on 3 benchmarks against 6. Pick Claude Sonnet 5 for output tokens 1.5x cheaper ($10.00 vs $15.00 per million) and work measured by LMArena WebDev or AA Coding Index, where it leads.

Which is cheaper, Claude Sonnet 5 or GPT-5.4?

Claude Sonnet 5 costs $2.00 per million input tokens and $10.00 per million output tokens; GPT-5.4 costs $2.50 and $15.00. Per solved SWE-bench Verified task, Claude Sonnet 5 costs $1.87 and GPT-5.4 $1.02.

Which has the bigger context window, Claude Sonnet 5 or GPT-5.4?

GPT-5.4, with 1.05M tokens against 1M.

How many benchmarks is the Claude Sonnet 5 vs GPT-5.4 comparison based on?

91 benchmarks have a result for both models as of October 21, 2026. Claude Sonnet 5's coding index rests on 6 independent benchmarks and GPT-5.4's on 3. 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-5-vs-gpt-5-4" title="Claude Sonnet 5 vs GPT-5.4, 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-5-vs-gpt-5-4">LLMBenchmark.ai, Claude Sonnet 5 vs GPT-5.4</a></p>