Enter opens the highlighted result. Esc closes.

Claude Opus 4.8 vs DeepSeek V3.2

Claude Opus 4.8 leads DeepSeek V3.2 on coding, 65.0 to 30.0 on the Vibe Coding Index, and DeepSeek V3.2 costs 28.4x less per output token ($0.88 vs $25.00 per million). Updated October 22, 2026; 62 benchmarks have a result for both.

Replace Claude Opus 4.8 Replace DeepSeek V3.2
Claude Opus 4.8Anthropic
Coding
65.0
Overall
69.8
Output $/M
$25.00
62
DeepSeek V3.2DeepSeek
Coding
30.0
Overall
37.9
Output $/M
$0.88
0
vs boards won of 62 shared

Claude Opus 4.8 or DeepSeek V3.2: which to pick

Pick Claude Opus 4.8 for coding: it scores 65.0 on the Vibe Coding Index against 30.0 for DeepSeek V3.2.

Pick DeepSeek V3.2 for output tokens 28.4x cheaper ($0.88 vs $25.00 per million) and a lower cost per solved task ($0.99 against $2.17).

Model AClaude Opus 4.8Anthropic65.0/1006 independent benchmarks$2.17 per solved taskModel BDeepSeek V3.2DeepSeek30.0/1003 independent benchmarks$0.99 per solved task
Shared results
62
Claude Opus 4.8 only
68
DeepSeek V3.2 only
13
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 Opus 4.8 65.0 from 6 benchmarks, DeepSeek V3.2 30.0 from 3 benchmarks.
Widest category
Coding, 26.2 points apart on average over 10 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): Claude Opus 4.8 82.7% against DeepSeek V3.2 5.1%, 77.6 points apart on a 0 to 100 scale.
Output price
Claude Opus 4.8 $25.00 against DeepSeek V3.2 $0.88 per million output tokens, 28.4x apart.
Context window
Claude Opus 4.8 takes 1M tokens, DeepSeek V3.2 164K.

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 4.8BarsDeepSeek V3.2Ties
Coding100
Intelligence20
Agentic work30
Reasoning80
Math20
Knowledge80
Web design80
General210
All shared boards620
CodingLike-for-likeSix of these boards feed the Vibe Coding Index100
Coding benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
SWE-bench VerifiedNot measured70.0%Only DeepSeek V3.2 measured
SWE-bench MultilingualNot measured59.0%Only DeepSeek V3.2 measured
LiveBench coding77.3%Not measuredOnly Claude Opus 4.8 measured
LMArena WebDev15561362Claude Opus 4.8
AA Coding Index74.344.2Claude Opus 4.8
LiveCodeBench (Vals AI)87.8%80.7%Claude Opus 4.8
SWE-bench Verified (Vals AI)88.6%67.6%Claude Opus 4.8
Terminal-Bench 4 (Vals AI)23.2%Not measuredOnly Claude Opus 4.8 measured
Vibe Code Bench (Vals AI)82.7%5.1%Claude Opus 4.8
Code migration (Vals AI)47.3%Not measuredOnly Claude Opus 4.8 measured
ProgramBench (Vals AI)1.0%Not measuredOnly Claude Opus 4.8 measured
Terminal-Bench 2.1 (Vals AI)71.9%Not measuredOnly Claude Opus 4.8 measured
SkillsBench (Vals AI)59.2%Not measuredOnly Claude Opus 4.8 measured
LiveCodeBench (Artificial Analysis run)Not measured86.2%Only DeepSeek V3.2 measured
SciCode54.4%Not measuredOnly Claude Opus 4.8 measured
Terminal-Bench Hard (Artificial Analysis run)58.3%35.6%Claude Opus 4.8
Terminal-BenchNot measured39.6%Only DeepSeek V3.2 measured
FrontierCode46.5%Not measuredOnly Claude Opus 4.8 measured
GSO-Bench47.1%Not measuredOnly Claude Opus 4.8 measured
WeirdML82.9%Not measuredOnly Claude Opus 4.8 measured
LMArena Text: coding14901453Claude Opus 4.8
LMArena Text: software and IT14881449Claude Opus 4.8
LMArena WebDev: React15571350Claude Opus 4.8
LMArena WebDev: HTML15431371Claude Opus 4.8
SciCode12.3%Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 14 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
AA Intelligence Index41.821.5Claude Opus 4.8
Epoch Capabilities Index159.0146.1Claude Opus 4.8
Agentic workLike-for-likePublic boards, shown as published, kept out of the index30
Agentic work benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
AA Agentic Index41.9Not measuredOnly Claude Opus 4.8 measured
Design Arena: agents building web apps1233Not measuredOnly Claude Opus 4.8 measured
Design Arena: agents building full-stack apps1249Not measuredOnly Claude Opus 4.8 measured
Design Arena: agents building mobile apps1222Not measuredOnly Claude Opus 4.8 measured
Design Arena: agents building Android apps1233Not measuredOnly Claude Opus 4.8 measured
Design Arena: agents building Godot games1231Not measuredOnly Claude Opus 4.8 measured
Design Arena: agents building games1239Not measuredOnly Claude Opus 4.8 measured
Design Arena: HTML slides1171Not measuredOnly Claude Opus 4.8 measured
Design Arena: PowerPoint slides1202Not measuredOnly Claude Opus 4.8 measured
Design Arena: spreadsheets1303Not measuredOnly Claude Opus 4.8 measured
LiveBench research81.8%Not measuredOnly Claude Opus 4.8 measured
LiveBench browser96.0%Not measuredOnly Claude Opus 4.8 measured
LiveBench long agentic48.3%Not measuredOnly Claude Opus 4.8 measured
Finance Agent (Vals AI)53.9%Not measuredOnly Claude Opus 4.8 measured
Tax Agent Bench (Vals AI)30.0%Not measuredOnly Claude Opus 4.8 measured
Legal Research Bench (Vals AI)43.8%Not measuredOnly Claude Opus 4.8 measured
Excel Modeling Benchmark (Vals AI)69.4%Not measuredOnly Claude Opus 4.8 measured
Terminal-Bench-Science (Vals AI)4.3%Not measuredOnly Claude Opus 4.8 measured
Time Horizon Index (Vals AI)11.8%Not measuredOnly Claude Opus 4.8 measured
τ²-Bench Telecom94.4%90.6%Claude Opus 4.8
APEX-Agents48.9%21.3%Claude Opus 4.8
Remote Labor Index8.3%Not measuredOnly Claude Opus 4.8 measured
OSWorld 2.020.6%Not measuredOnly Claude Opus 4.8 measured
DeepResearch Bench50.2%Not measuredOnly Claude Opus 4.8 measured
PostTrainBench33.8%Not measuredOnly Claude Opus 4.8 measured
Vending-Bench 2$5,787$1,034Claude Opus 4.8
LMArena Search1204Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 14 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index80
Reasoning benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
GPQA Diamond92.4%80.3%Claude Opus 4.8
LiveBench reasoning89.2%Not measuredOnly Claude Opus 4.8 measured
MMMU Pro (Vals AI)86.6%Not measuredOnly Claude Opus 4.8 measured
GPQA Diamond (Artificial Analysis run)92.0%84.0%Claude Opus 4.8
Humanity's Last Exam (Artificial Analysis run)48.7%24.6%Claude Opus 4.8
GPQA Diamond (Epoch AI run)91.0%Not measuredOnly Claude Opus 4.8 measured
ARC-AGI92.5%57.0%Claude Opus 4.8
ARC-AGI-272.1%4.0%Claude Opus 4.8
SimpleBench64.8%Not measuredOnly Claude Opus 4.8 measured
Chess puzzles52.0%Not measuredOnly Claude Opus 4.8 measured
Mystery game puzzles36.0%Not measuredOnly Claude Opus 4.8 measured
EBR-bench28.6%Not measuredOnly Claude Opus 4.8 measured
LMArena Text: hard prompts14821434Claude Opus 4.8
LMArena Text: expert15031436Claude Opus 4.8
LMArena Text: hard prompts in English14871446Claude Opus 4.8
LMArena Vision: diagrams1315Not measuredOnly Claude Opus 4.8 measured
LMArena Vision: homework1337Not measuredOnly Claude Opus 4.8 measured
Kagi LLM Benchmark88.8%Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
AIMENot measured84.6%Only DeepSeek V3.2 measured
MGSM (Vals AI)Not measured86.0%Only DeepSeek V3.2 measured
AA Math IndexNot measured92.0Only DeepSeek V3.2 measured
AIME 2025 (Artificial Analysis run)Not measured92.0%Only DeepSeek V3.2 measured
OTIS Mock AIME98.3%Not measuredOnly Claude Opus 4.8 measured
FrontierMath, tiers 1 to 380.0%Not measuredOnly Claude Opus 4.8 measured
FrontierMath, tier 456.1%Not measuredOnly Claude Opus 4.8 measured
ProofBench69.0%Not measuredOnly Claude Opus 4.8 measured
LMArena Text: math14861435Claude Opus 4.8
LMArena Text: mathematical work14971440Claude Opus 4.8
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index80
Knowledge benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
MMLU Pro89.6%84.9%Claude Opus 4.8
LegalBench (Vals AI)83.6%76.1%Claude Opus 4.8
CaseLaw (Vals AI)Not measured55.4%Only DeepSeek V3.2 measured
CorpFin (Vals AI)66.7%51.0%Claude Opus 4.8
TaxEval (Vals AI)75.6%68.2%Claude Opus 4.8
MortgageTax (Vals AI)69.9%Not measuredOnly Claude Opus 4.8 measured
MedQA (Vals AI)Not measured93.9%Only DeepSeek V3.2 measured
MedCode (Vals AI)53.2%Not measuredOnly Claude Opus 4.8 measured
MMLU Pro (Artificial Analysis run)Not measured86.2%Only DeepSeek V3.2 measured
SimpleQA Verified53.0%Not measuredOnly Claude Opus 4.8 measured
LMArena Text: science14771446Claude Opus 4.8
LMArena Text: medicine and healthcare14641440Claude Opus 4.8
LMArena Text: legal and government14771425Claude Opus 4.8
LMArena Text: business and finance14581410Claude Opus 4.8
FACTS Benchmark Suite53.5%Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 3 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
Design Arena: websites12581175Claude Opus 4.8
Design Arena: UI components12551146Claude Opus 4.8
Design Arena: game dev12641131Claude Opus 4.8
Design Arena: data viz12401158Claude Opus 4.8
Design Arena: 3D12221132Claude Opus 4.8
Design Arena: all code categories12551163Claude Opus 4.8
Design Arena: SVG11951032Claude Opus 4.8
Design Arena: ASCII art12771083Claude Opus 4.8
LiveBench design78.8%Not measuredOnly Claude Opus 4.8 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index210
General benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
LiveBench data analysis66.0%Not measuredOnly Claude Opus 4.8 measured
Public Benefits Bench (Vals AI)68.1%Not measuredOnly Claude Opus 4.8 measured
SAGE (Vals AI)54.8%Not measuredOnly Claude Opus 4.8 measured
MedScribe (Vals AI)85.8%Not measuredOnly Claude Opus 4.8 measured
Vals Multimodal Index70.9%Not measuredOnly Claude Opus 4.8 measured
IFBench62.2%60.7%Claude Opus 4.8
Long context reasoning (AA-LCR)77.7%73.3%Claude Opus 4.8
DTBench94.9%85.6%Claude Opus 4.8
LMCA57.5%28.8%Claude Opus 4.8
CL-benchNot measured12.4%Only DeepSeek V3.2 measured
Surface Evolver Bench87.5%Not measuredOnly Claude Opus 4.8 measured
Furniture assembly42.5%Not measuredOnly Claude Opus 4.8 measured
CL-bench LifeNot measured7.4%Only DeepSeek V3.2 measured
LMArena Text14611425Claude Opus 4.8
LMArena Text: instruction following14761413Claude Opus 4.8
LMArena Text: creative writing14541400Claude Opus 4.8
LMArena Vision1294Not measuredOnly Claude Opus 4.8 measured
LMArena Text: English14681439Claude Opus 4.8
LMArena Text: non-English14501409Claude Opus 4.8
LMArena Text: Chinese15061455Claude Opus 4.8
LMArena Text: German14721413Claude Opus 4.8
LMArena Text: Japanese14421374Claude Opus 4.8
LMArena Text: Korean14311370Claude Opus 4.8
LMArena Text: French14811429Claude Opus 4.8
LMArena Text: Spanish14651424Claude Opus 4.8
LMArena Text: Russian14741424Claude Opus 4.8
LMArena Text: Polish14591419Claude Opus 4.8
LMArena Text: multi-turn14761427Claude Opus 4.8
LMArena Text: longer queries14831428Claude Opus 4.8
LMArena Text: writing and language14611407Claude Opus 4.8
LMArena Text: entertainment and media14401395Claude Opus 4.8
LMArena Vision: English1293Not measuredOnly Claude Opus 4.8 measured
LMArena Vision: Chinese1346Not measuredOnly Claude Opus 4.8 measured
LMArena Vision: reading text1303Not measuredOnly Claude Opus 4.8 measured
LMArena Vision: humor1282Not measuredOnly Claude Opus 4.8 measured
LMArena Vision: creative writing1301Not measuredOnly Claude Opus 4.8 measured
LMArena Document1475Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 9 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 4.8 and DeepSeek V3.2
BenchmarkClaude Opus 4.8DeepSeek V3.2Leader
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 4.8$0.018Fits in one request
DeepSeek V3.2$0.0007Fits in one request

DeepSeek V3.2 costs less

Repository review

50K input + 3K output tokens

Claude Opus 4.8$0.325Fits in one request
DeepSeek V3.2$0.017Fits in one request

DeepSeek V3.2 costs less

Agent loop

200K input + 10K output tokens

Claude Opus 4.8$1.25Fits in one request
DeepSeek V3.2$0.067Exceeds the context window

DeepSeek V3.2 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 4.8 and DeepSeek V3.2
SpecificationAnthropic Claude Opus 4.8DeepSeek V3.2
MakerAnthropicDeepSeek
WeightsNo public weights listedOpen weights
ReleasedMay 27, 2026December 1, 2025
Context window1M164K
Input $/Mtok$5.00$0.29
Output $/Mtok$25.00$0.88
Coding index65.030.0
Independent measurements behind it63
Cost per solved task$2.17$0.99
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 4.8 better than DeepSeek V3.2 for coding?

Pick Claude Opus 4.8 for coding: it scores 65.0 on the Vibe Coding Index against 30.0 for DeepSeek V3.2. Pick DeepSeek V3.2 for output tokens 28.4x cheaper ($0.88 vs $25.00 per million) and a lower cost per solved task ($0.99 against $2.17).

Which is cheaper, Claude Opus 4.8 or DeepSeek V3.2?

Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; DeepSeek V3.2 costs $0.29 and $0.88. Per solved SWE-bench Verified task, Claude Opus 4.8 costs $2.17 and DeepSeek V3.2 $0.99.

Which has the bigger context window, Claude Opus 4.8 or DeepSeek V3.2?

Claude Opus 4.8, with 1M tokens against 164K.

How many benchmarks is the Claude Opus 4.8 vs DeepSeek V3.2 comparison based on?

62 benchmarks have a result for both models as of October 22, 2026. Claude Opus 4.8's coding index rests on 6 independent benchmarks and DeepSeek V3.2'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-4-8-vs-deepseek-v3-2" title="Claude Opus 4.8 vs DeepSeek V3.2, 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-4-8-vs-deepseek-v3-2">LLMBenchmark.ai, Claude Opus 4.8 vs DeepSeek V3.2</a></p>