LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Opus 5.5 vs DeepSeek V4 Flash 0423

Updated October 21, 2026. 40 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Claude Opus 5.5 Replace DeepSeek V4 Flash 0423
Claude Opus 5.5Anthropic
Coding
92.2
Overall
57.6
Output $/M
$20.00
40
DeepSeek V4 Flash 0423DeepSeek
Coding
54.6
Overall
24.4
Output $/M
$0.182
0
vs boards won of 40 shared

Decision reading

Pick Claude Opus 5.5 for coding: it scores 92.2 on the Vibe Coding Index against 54.6 for DeepSeek V4 Flash 0423.

Pick DeepSeek V4 Flash 0423 when output tokens cost 109.9x less, or when the context window is larger, 1.05M against 1M tokens.

Model AClaude Opus 5.5Anthropic92.2/1002 independent benchmarksNo cost per solved taskModel BDeepSeek V4 Flash 0423DeepSeek54.6/1001 independent benchmarkNo cost per solved task
Shared results
40
Claude Opus 5.5 only
67
DeepSeek V4 Flash 0423 only
8
Like-for-like categories
7 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

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 5.5 92.2 from 2 benchmarks, DeepSeek V4 Flash 0423 54.6 from 1 benchmark.
Widest category
Intelligence, 30.8 points apart on average over 2 shared benchmarks.
Widest single gap
ALE-Bench: Claude Opus 5.5 2147.4 against DeepSeek V4 Flash 0423 678.2, 49.0 points apart on a 0 to 100 scale.
Output price
Claude Opus 5.5 $20.00 against DeepSeek V4 Flash 0423 $0.182 per million output tokens, 109.9x apart.
Context window
Claude Opus 5.5 takes 1M tokens, DeepSeek V4 Flash 0423 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 Opus 5.5BarsDeepSeek V4 Flash 0423Ties
Coding60
Intelligence20
Reasoning30
Math20
Knowledge40
Web design70
General160
All shared boards400
CodingLike-for-likeFour of these boards feed the Vibe Coding Index60
Coding benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
LiveBench coding96.3%Not measuredOnly Claude Opus 5.5 measured
LMArena WebDev18071582Claude Opus 5.5
AA Coding IndexNot measured52.0Only DeepSeek V4 Flash 0423 measured
Terminal-Bench 4 (Vals AI)65.2%Not measuredOnly Claude Opus 5.5 measured
Vibe Code Bench (Vals AI)90.3%Not measuredOnly Claude Opus 5.5 measured
Code migration (Vals AI)66.7%Not measuredOnly Claude Opus 5.5 measured
ProgramBench (Vals AI)18.5%Not measuredOnly Claude Opus 5.5 measured
IOI (Vals AI)95.1%Not measuredOnly Claude Opus 5.5 measured
Terminal-Bench 2.1 (Vals AI)87.6%Not measuredOnly Claude Opus 5.5 measured
SWE-bench Multimodal (Vals AI)70.2%Not measuredOnly Claude Opus 5.5 measured
Vibe Code Bench 1-100 (Vals AI)30.4%Not measuredOnly Claude Opus 5.5 measured
CyberBench (Vals AI)74.6%Not measuredOnly Claude Opus 5.5 measured
SciCode66.9%Not measuredOnly Claude Opus 5.5 measured
FrontierCode54.6%Not measuredOnly Claude Opus 5.5 measured
FrontierSWE62.3%Not measuredOnly Claude Opus 5.5 measured
WeirdMLNot measured45.6%Only DeepSeek V4 Flash 0423 measured
MirrorCode77.4%Not measuredOnly Claude Opus 5.5 measured
ALE-Bench2147.4678.2Claude Opus 5.5
LMArena Text: coding15451457Claude Opus 5.5
LMArena Text: software and IT15381453Claude Opus 5.5
LMArena WebDev: React18551585Claude Opus 5.5
LMArena WebDev: HTML17531558Claude Opus 5.5
LMArena WebDev: from an image1753Not measuredOnly Claude Opus 5.5 measured
Neither model has a result on the other 15 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
AA Intelligence Index57.624.4Claude Opus 5.5
Epoch Capabilities Index168.8146.1Claude Opus 5.5
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index24 boards, none shared
Agentic work benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
AA Agentic IndexNot measured26.3Only DeepSeek V4 Flash 0423 measured
Design Arena: agents building web apps1316Not measuredOnly Claude Opus 5.5 measured
Design Arena: agents building mobile apps1311Not measuredOnly Claude Opus 5.5 measured
Design Arena: spreadsheets1359Not measuredOnly Claude Opus 5.5 measured
LiveBench research85.0%Not measuredOnly Claude Opus 5.5 measured
LiveBench browser80.0%Not measuredOnly Claude Opus 5.5 measured
LiveBench long agentic85.0%Not measuredOnly Claude Opus 5.5 measured
SREBench (Vals AI)33.6%Not measuredOnly Claude Opus 5.5 measured
Finance Agent (Vals AI)58.6%Not measuredOnly Claude Opus 5.5 measured
Tax Agent Bench (Vals AI)45.6%Not measuredOnly Claude Opus 5.5 measured
Legal Research Bench (Vals AI)50.5%Not measuredOnly Claude Opus 5.5 measured
Harvey's Legal Agent Benchmark (Vals AI)9.2%Not measuredOnly Claude Opus 5.5 measured
Excel Modeling Benchmark (Vals AI)75.9%Not measuredOnly Claude Opus 5.5 measured
BioMysteryBench (Vals AI)79.3%Not measuredOnly Claude Opus 5.5 measured
Terminal-Bench-Science (Vals AI)47.1%Not measuredOnly Claude Opus 5.5 measured
Vals Science Index58.7%Not measuredOnly Claude Opus 5.5 measured
Time Horizon Index (Vals AI)91.3%Not measuredOnly Claude Opus 5.5 measured
CUA-bench (Vals AI)14.0%Not measuredOnly Claude Opus 5.5 measured
Robo-bench (Vals AI)14.6%Not measuredOnly Claude Opus 5.5 measured
APEX-Agents73.5%Not measuredOnly Claude Opus 5.5 measured
Remote Labor Index21.3%Not measuredOnly Claude Opus 5.5 measured
SpireClimb3.3%Not measuredOnly Claude Opus 5.5 measured
Vending-Bench 2$9,235Not measuredOnly Claude Opus 5.5 measured
LMArena Search1213Not measuredOnly Claude Opus 5.5 measured
Neither model has a result on the other 15 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index30
Reasoning benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
LiveBench reasoning92.2%Not measuredOnly Claude Opus 5.5 measured
MysteryMechanism (Vals AI)49.5%Not measuredOnly Claude Opus 5.5 measured
Humanity's Last Exam (Artificial Analysis run)61.4%Not measuredOnly Claude Opus 5.5 measured
GPQA Diamond (Epoch AI run)90.6%Not measuredOnly Claude Opus 5.5 measured
ARC-AGI98.5%Not measuredOnly Claude Opus 5.5 measured
ARC-AGI-293.3%Not measuredOnly Claude Opus 5.5 measured
SimpleBenchNot measured46.3%Only DeepSeek V4 Flash 0423 measured
Chess puzzles67.0%Not measuredOnly Claude Opus 5.5 measured
Mystery game puzzles71.0%Not measuredOnly Claude Opus 5.5 measured
EBR-bench71.4%Not measuredOnly Claude Opus 5.5 measured
LMArena Text: hard prompts15361444Claude Opus 5.5
LMArena Text: expert15491441Claude Opus 5.5
LMArena Text: hard prompts in English15351449Claude Opus 5.5
LMArena Vision: diagrams1345Not measuredOnly Claude Opus 5.5 measured
LMArena Vision: homework1362Not measuredOnly Claude Opus 5.5 measured
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
OTIS Mock AIME100.0%Not measuredOnly Claude Opus 5.5 measured
FrontierMath, tiers 1 to 391.2%Not measuredOnly Claude Opus 5.5 measured
FrontierMath, tier 495.0%Not measuredOnly Claude Opus 5.5 measured
ProofBench100.0%Not measuredOnly Claude Opus 5.5 measured
LMArena Text: math15181426Claude Opus 5.5
LMArena Text: mathematical work15161432Claude Opus 5.5
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
Knowledge benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
MedCode (Vals AI)49.8%Not measuredOnly Claude Opus 5.5 measured
SimpleQA Verified72.2%Not measuredOnly Claude Opus 5.5 measured
LMArena Text: science15241446Claude Opus 5.5
LMArena Text: medicine and healthcare15091448Claude Opus 5.5
LMArena Text: legal and government15181439Claude Opus 5.5
LMArena Text: business and finance14941425Claude Opus 5.5
Neither model has a result on the other 11 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index70
Web design benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
Design Arena: websites13351208Claude Opus 5.5
Design Arena: UI components13901164Claude Opus 5.5
Design Arena: game dev14571197Claude Opus 5.5
Design Arena: data viz13971127Claude Opus 5.5
Design Arena: 3D14771189Claude Opus 5.5
Design Arena: all code categories13901205Claude Opus 5.5
Design Arena: SVG13321155Claude Opus 5.5
Design Arena: ASCII artNot measured1109Only DeepSeek V4 Flash 0423 measured
LiveBench design86.9%Not measuredOnly Claude Opus 5.5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index160
General benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
Vals Index70.2%Not measuredOnly Claude Opus 5.5 measured
LiveBench data analysis80.3%Not measuredOnly Claude Opus 5.5 measured
Public Benefits Bench (Vals AI)70.6%Not measuredOnly Claude Opus 5.5 measured
SAGE (Vals AI)45.8%Not measuredOnly Claude Opus 5.5 measured
MedScribe (Vals AI)91.4%Not measuredOnly Claude Opus 5.5 measured
Long context reasoning (AA-LCR)84.7%Not measuredOnly Claude Opus 5.5 measured
DTBench98.9%86.4%Claude Opus 5.5
LMCA68.2%35.9%Claude Opus 5.5
Furniture assembly83.3%Not measuredOnly Claude Opus 5.5 measured
LMArena Text15161432Claude Opus 5.5
LMArena Text: instruction following15371421Claude Opus 5.5
LMArena Text: creative writing15351403Claude Opus 5.5
LMArena Vision1321Not measuredOnly Claude Opus 5.5 measured
LMArena Text: English15191441Claude Opus 5.5
LMArena Text: non-English15071419Claude Opus 5.5
LMArena Text: Chinese15821463Claude Opus 5.5
LMArena Text: GermanNot measured1419Only DeepSeek V4 Flash 0423 measured
LMArena Text: Japanese15341389Claude Opus 5.5
LMArena Text: KoreanNot measured1383Only DeepSeek V4 Flash 0423 measured
LMArena Text: French14931439Claude Opus 5.5
LMArena Text: Spanish15081436Claude Opus 5.5
LMArena Text: Russian15171427Claude Opus 5.5
LMArena Text: PolishNot measured1436Only DeepSeek V4 Flash 0423 measured
LMArena Text: multi-turn15011449Claude Opus 5.5
LMArena Text: longer queries15321434Claude Opus 5.5
LMArena Text: writing and language15321410Claude Opus 5.5
LMArena Text: entertainment and media15151400Claude Opus 5.5
LMArena Vision: English1317Not measuredOnly Claude Opus 5.5 measured
LMArena Vision: reading text1326Not measuredOnly Claude Opus 5.5 measured
LMArena Vision: creative writing1320Not measuredOnly Claude Opus 5.5 measured
Neither model has a result on the other 16 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 5.5 and DeepSeek V4 Flash 0423
BenchmarkClaude Opus 5.5DeepSeek V4 Flash 0423Leader
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 5.5$0.014Fits in one request
DeepSeek V4 Flash 0423$0.0002Fits in one request

DeepSeek V4 Flash 0423 costs less

Repository review

50K input + 3K output tokens

Claude Opus 5.5$0.260Fits in one request
DeepSeek V4 Flash 0423$0.0051Fits in one request

DeepSeek V4 Flash 0423 costs less

Agent loop

200K input + 10K output tokens

Claude Opus 5.5$1.00Fits in one request
DeepSeek V4 Flash 0423$0.020Fits in one request

DeepSeek V4 Flash 0423 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 5.5 and DeepSeek V4 Flash 0423
SpecificationAnthropic Claude Opus 5.5DeepSeek V4 Flash 0423
MakerAnthropicDeepSeek
WeightsNo public weights listedOpen weights
ReleasedSeptember 22, 2026April 24, 2026
Context window1M1.05M
Input $/Mtok$4.00$0.091
Output $/Mtok$20.00$0.182
Coding index92.254.6
Independent measurements behind it21
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 5.5 better than DeepSeek V4 Flash 0423 for coding?

Pick Claude Opus 5.5 for coding: it scores 92.2 on the Vibe Coding Index against 54.6 for DeepSeek V4 Flash 0423. Pick DeepSeek V4 Flash 0423 when output tokens cost 109.9x less, or when the context window is larger, 1.05M against 1M tokens.

Which is cheaper, Claude Opus 5.5 or DeepSeek V4 Flash 0423?

Claude Opus 5.5 costs $4.00 per million input tokens and $20.00 per million output tokens; DeepSeek V4 Flash 0423 costs $0.091 and $0.182. Cost per solved task is known for neither of them.

Which has the bigger context window, Claude Opus 5.5 or DeepSeek V4 Flash 0423?

DeepSeek V4 Flash 0423, with 1.05M tokens against 1M.

How many benchmarks is the Claude Opus 5.5 vs DeepSeek V4 Flash 0423 comparison based on?

40 benchmarks have a result for both models as of October 21, 2026. Claude Opus 5.5's coding index rests on 2 independent benchmarks and DeepSeek V4 Flash 0423's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next