Enter opens the highlighted result. Esc closes.

Claude Opus 4.7 vs GPT-5.2

Claude Opus 4.7 leads GPT-5.2 on coding, 67.5 to 55.9 on the Vibe Coding Index, and GPT-5.2 costs 1.8x less per output token ($14.00 vs $25.00 per million). Updated October 21, 2026; 84 benchmarks have a result for both.

Replace Claude Opus 4.7 Replace GPT-5.2
Claude Opus 4.7Anthropic
Coding
67.5
Overall
71.6
Output $/M
$25.00
71
GPT-5.2OpenAI
Coding
55.9
Overall
60.7
Output $/M
$14.00
13
vs boards won of 84 shared

Claude Opus 4.7 or GPT-5.2: which to pick

Pick Claude Opus 4.7 for coding: it scores 67.5 on the Vibe Coding Index against 55.9 for GPT-5.2.

Pick GPT-5.2 for output tokens 1.8x cheaper ($14.00 vs $25.00 per million) and a lower cost per solved task ($1.91 against $2.95).

Model AClaude Opus 4.7Anthropic67.5/1004 independent benchmarks$2.95 per solved taskModel BGPT-5.2OpenAI55.9/1003 independent benchmarks$1.91 per solved task
Shared results
84
Claude Opus 4.7 only
19
GPT-5.2 only
30
Like-for-like categories
7 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 Opus 4.7 67.5 from 4 benchmarks, GPT-5.2 55.9 from 3 benchmarks.
Widest category
Agentic work, 24.9 points apart on average over 2 shared benchmarks.
Widest single gap
Vending-Bench 2: Claude Opus 4.7 $10,937 against GPT-5.2 $3,591, 45.9 points apart on a 0 to 100 scale.
Output price
Claude Opus 4.7 $25.00 against GPT-5.2 $14.00 per million output tokens, 1.8x apart.
Context window
Claude Opus 4.7 takes 1M tokens, GPT-5.2 400K.

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 Opus 4.7BarsGPT-5.2Ties
Coding101
Intelligence20
Agentic work20
Reasoning125
Math61
Knowledge112
General284
All shared boards7113
CodingLike-for-likeSix of these boards feed the Vibe Coding Index101
Coding benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
SWE-bench VerifiedNot measured72.8%Only GPT-5.2 measured
SWE-bench MultilingualNot measured66.7%Only GPT-5.2 measured
LMArena WebDev15581417Claude Opus 4.7
AA Coding Index73.6Not measuredOnly Claude Opus 4.7 measured
LiveCodeBench (Vals AI)85.1%85.4%GPT-5.2
SWE-bench Verified (Vals AI)82.0%75.8%Claude Opus 4.7
Vibe Code Bench (Vals AI)71.0%53.5%Claude Opus 4.7
Code migration (Vals AI)43.9%Not measuredOnly Claude Opus 4.7 measured
ProgramBench (Vals AI)0.0%Not measuredOnly Claude Opus 4.7 measured
Terminal-Bench 2.1 (Vals AI)68.5%Not measuredOnly Claude Opus 4.7 measured
LiveCodeBench (Artificial Analysis run)Not measured89.4%Only GPT-5.2 measured
Terminal-Bench Hard (Artificial Analysis run)54.5%47.0%Claude Opus 4.7
Terminal-Bench80.2%64.9%Claude Opus 4.7
FrontierCode38.5%Not measuredOnly Claude Opus 4.7 measured
GSO-Bench44.1%27.5%Claude Opus 4.7
WeirdML76.4%72.2%Claude Opus 4.7
MirrorCode31.1%Not measuredOnly Claude Opus 4.7 measured
ALE-BenchNot measured1293.5Only GPT-5.2 measured
LMArena Text: coding15191442Claude Opus 4.7
LMArena Text: software and IT15161434Claude Opus 4.7
LMArena WebDev: React1560Not measuredOnly Claude Opus 4.7 measured
LMArena WebDev: HTML15541427Claude Opus 4.7
LMArena WebDev: from an image1577Not measuredOnly Claude Opus 4.7 measured
Neither model has a result on the other 16 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
AA Intelligence Index40.730.4Claude Opus 4.7
Epoch Capabilities Index156.8153.7Claude Opus 4.7
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
Agentic work benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
AA Agentic Index38.6Not measuredOnly Claude Opus 4.7 measured
Design Arena: agents building web appsNot measured1086Only GPT-5.2 measured
Design Arena: agents building full-stack appsNot measured1043Only GPT-5.2 measured
Design Arena: agents building mobile appsNot measured1093Only GPT-5.2 measured
Design Arena: agents building Android appsNot measured1058Only GPT-5.2 measured
Design Arena: agents building Godot gamesNot measured1142Only GPT-5.2 measured
Finance Agent (Vals AI)51.5%Not measuredOnly Claude Opus 4.7 measured
Tax Agent Bench (Vals AI)23.7%Not measuredOnly Claude Opus 4.7 measured
Legal Research Bench (Vals AI)38.5%Not measuredOnly Claude Opus 4.7 measured
Excel Modeling Benchmark (Vals AI)63.7%Not measuredOnly Claude Opus 4.7 measured
τ²-Bench Telecom88.6%84.8%Claude Opus 4.7
APEX-Agents49.2%Not measuredOnly Claude Opus 4.7 measured
METR time horizonsNot measured75.3%Only GPT-5.2 measured
Remote Labor IndexNot measured2.5%Only GPT-5.2 measured
GDPvalNot measured49.7%Only GPT-5.2 measured
OSWorld 2.018.2%Not measuredOnly Claude Opus 4.7 measured
DeepResearch BenchNot measured41.1%Only GPT-5.2 measured
PostTrainBench28.6%Not measuredOnly Claude Opus 4.7 measured
Vending-Bench 2$10,937$3,591Claude Opus 4.7
LMArena Search1230Not measuredOnly Claude Opus 4.7 measured
Berkeley Function Calling Leaderboard (BFCL)Not measured55.9%Only GPT-5.2 measured
Neither model has a result on the other 20 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index125
Reasoning benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
GPQA Diamond90.2%91.7%GPT-5.2
MMMU Pro (Vals AI)85.5%86.7%GPT-5.2
GPQA Diamond (Artificial Analysis run)91.4%90.3%Claude Opus 4.7
Humanity's Last Exam (Artificial Analysis run)42.3%37.7%Claude Opus 4.7
GPQA Diamond (Epoch AI run)90.2%91.4%GPT-5.2
ARC-AGI93.5%86.2%Claude Opus 4.7
ARC-AGI-275.8%52.9%Claude Opus 4.7
SimpleBench61.7%45.8%Claude Opus 4.7
Chess puzzles41.0%48.0%GPT-5.2
Mystery game puzzles28.0%23.0%Claude Opus 4.7
EBR-bench19.0%23.0%GPT-5.2
VPCTNot measured84.0%Only GPT-5.2 measured
LMArena Text: hard prompts15061428Claude Opus 4.7
LMArena Text: expert15211445Claude Opus 4.7
LMArena Text: hard prompts in English15081431Claude Opus 4.7
LMArena Vision: diagrams13361265Claude Opus 4.7
LMArena Vision: homework13371276Claude Opus 4.7
Kagi LLM Benchmark80.7%73.3%Claude Opus 4.7
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index61
Math benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
AIME96.3%96.9%GPT-5.2
MGSM (Vals AI)Not measured94.0%Only GPT-5.2 measured
AA Math IndexNot measured99.0Only GPT-5.2 measured
AIME 2025 (Artificial Analysis run)Not measured99.0%Only GPT-5.2 measured
OTIS Mock AIME97.8%96.1%Claude Opus 4.7
FrontierMath, tiers 1 to 370.2%67.4%Claude Opus 4.7
FrontierMath, tier 431.7%31.7%Claude Opus 4.7
ProofBench54.0%15.0%Claude Opus 4.7
LMArena Text: math14981440Claude Opus 4.7
LMArena Text: mathematical work15091442Claude Opus 4.7
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index112
Knowledge benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
MMLU Pro89.9%86.2%Claude Opus 4.7
LegalBench (Vals AI)85.3%82.8%Claude Opus 4.7
CaseLaw (Vals AI)68.4%66.0%Claude Opus 4.7
CorpFin (Vals AI)66.1%65.9%Claude Opus 4.7
TaxEval (Vals AI)75.3%75.8%GPT-5.2
MortgageTax (Vals AI)70.3%67.1%Claude Opus 4.7
MedQA (Vals AI)Not measured94.1%Only GPT-5.2 measured
MedCode (Vals AI)54.9%49.8%Claude Opus 4.7
MMLU Pro (Artificial Analysis run)Not measured87.4%Only GPT-5.2 measured
SimpleQA Verified51.7%37.1%Claude Opus 4.7
LMArena Text: science15081425Claude Opus 4.7
LMArena Text: medicine and healthcare14951425Claude Opus 4.7
LMArena Text: legal and government15041427Claude Opus 4.7
LMArena Text: business and finance14901414Claude Opus 4.7
FACTS Benchmark Suite45.5%48.5%GPT-5.2
Neither model has a result on the other 3 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
Design Arena: websitesNot measured1195Only GPT-5.2 measured
Design Arena: UI componentsNot measured1190Only GPT-5.2 measured
Design Arena: game devNot measured1195Only GPT-5.2 measured
Design Arena: data vizNot measured1202Only GPT-5.2 measured
Design Arena: 3DNot measured1079Only GPT-5.2 measured
Design Arena: all code categoriesNot measured1170Only GPT-5.2 measured
Design Arena: SVGNot measured1139Only GPT-5.2 measured
Design Arena: ASCII artNot measured1152Only GPT-5.2 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index284
General benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
SAGE (Vals AI)56.1%49.3%Claude Opus 4.7
MedScribe (Vals AI)83.0%84.4%GPT-5.2
Vals Multimodal Index67.4%Not measuredOnly Claude Opus 4.7 measured
IFBench58.6%75.4%GPT-5.2
Long context reasoning (AA-LCR)78.7%82.7%GPT-5.2
DTBench94.7%90.9%Claude Opus 4.7
LMCA52.2%43.9%Claude Opus 4.7
CL-benchNot measured18.2%Only GPT-5.2 measured
Furniture assembly33.3%38.3%GPT-5.2
ExploitBench26.5%Not measuredOnly Claude Opus 4.7 measured
LMArena Text14901416Claude Opus 4.7
LMArena Text: instruction following14981409Claude Opus 4.7
LMArena Text: creative writing14861376Claude Opus 4.7
LMArena Vision13171246Claude Opus 4.7
LMArena Text: English14961422Claude Opus 4.7
LMArena Text: non-English14801405Claude Opus 4.7
LMArena Text: Chinese15311447Claude Opus 4.7
LMArena Text: German14951444Claude Opus 4.7
LMArena Text: Japanese14741420Claude Opus 4.7
LMArena Text: Korean14641374Claude Opus 4.7
LMArena Text: French15031454Claude Opus 4.7
LMArena Text: Spanish14941413Claude Opus 4.7
LMArena Text: Russian14941415Claude Opus 4.7
LMArena Text: Polish14891426Claude Opus 4.7
LMArena Text: multi-turn15051419Claude Opus 4.7
LMArena Text: longer queries15051413Claude Opus 4.7
LMArena Text: writing and language14921392Claude Opus 4.7
LMArena Text: entertainment and media14741378Claude Opus 4.7
LMArena Vision: English13221243Claude Opus 4.7
LMArena Vision: Chinese13711305Claude Opus 4.7
LMArena Vision: reading text13271257Claude Opus 4.7
LMArena Vision: humor13091227Claude Opus 4.7
LMArena Vision: captioningNot measured1267Only GPT-5.2 measured
LMArena Vision: recognizing things12781232Claude Opus 4.7
LMArena Vision: creative writing13261225Claude Opus 4.7
LMArena Document14951405Claude Opus 4.7
Neither model has a result on the other 10 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 4.7 and GPT-5.2
BenchmarkClaude Opus 4.7GPT-5.2Leader
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 Opus 4.7$0.018Fits in one request
GPT-5.2$0.0088Fits in one request

GPT-5.2 costs less

Repository review

50K input + 3K output tokens

Claude Opus 4.7$0.325Fits in one request
GPT-5.2$0.130Fits in one request

GPT-5.2 costs less

Agent loop

200K input + 10K output tokens

Claude Opus 4.7$1.25Fits in one request
GPT-5.2$0.490Fits in one request

GPT-5.2 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 Opus 4.7 and GPT-5.2
SpecificationAnthropic Claude Opus 4.7OpenAI GPT-5.2
MakerAnthropicOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedApril 16, 2026December 10, 2025
Context window1M400K
Input $/Mtok$5.00$1.75
Output $/Mtok$25.00$14.00
Coding index67.555.9
Independent measurements behind it43
Cost per solved task$2.95$1.91
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 4.7 better than GPT-5.2 for coding?

Pick Claude Opus 4.7 for coding: it scores 67.5 on the Vibe Coding Index against 55.9 for GPT-5.2. Pick GPT-5.2 for output tokens 1.8x cheaper ($14.00 vs $25.00 per million) and a lower cost per solved task ($1.91 against $2.95).

Which is cheaper, Claude Opus 4.7 or GPT-5.2?

Claude Opus 4.7 costs $5.00 per million input tokens and $25.00 per million output tokens; GPT-5.2 costs $1.75 and $14.00. Per solved SWE-bench Verified task, Claude Opus 4.7 costs $2.95 and GPT-5.2 $1.91.

Which has the bigger context window, Claude Opus 4.7 or GPT-5.2?

Claude Opus 4.7, with 1M tokens against 400K.

How many benchmarks is the Claude Opus 4.7 vs GPT-5.2 comparison based on?

84 benchmarks have a result for both models as of October 21, 2026. Claude Opus 4.7's coding index rests on 4 independent benchmarks and GPT-5.2'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-opus-4-7-vs-gpt-5-2" title="Claude Opus 4.7 vs GPT-5.2, 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-opus-4-7-vs-gpt-5-2">LLMBenchmark.ai, Claude Opus 4.7 vs GPT-5.2</a></p>