LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Haiku 4.5 vs DeepSeek V4 Pro 0423

Updated October 21, 2026. 70 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Claude Haiku 4.5 Replace DeepSeek V4 Pro 0423
Claude Haiku 4.5Anthropic
Coding
49.8
Overall
16.9
Output $/M
$5.00
2
DeepSeek V4 Pro 0423DeepSeek
Coding
35.3
Overall
30.4
Output $/M
$1.9105
67
vs boards won of 70 shared, 1 tied

Decision reading

Pick Claude Haiku 4.5 for coding: it scores 49.8 on the Vibe Coding Index against 35.3 for DeepSeek V4 Pro 0423.

Pick DeepSeek V4 Pro 0423 when output tokens cost 2.6x less, or when the context window is larger, 1.05M against 200K tokens.

Model AClaude Haiku 4.5Anthropic49.8/1003 independent benchmarks$0.56 per solved taskModel BDeepSeek V4 Pro 0423DeepSeek35.3/1001 independent benchmark$0.57 per solved task
Shared results
70
Claude Haiku 4.5 only
24
DeepSeek V4 Pro 0423 only
13
Like-for-like categories
8 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 Haiku 4.5 49.8 from 3 benchmarks, DeepSeek V4 Pro 0423 35.3 from 1 benchmark.
Widest category
Agentic work, 18.3 points apart on average over 6 shared benchmarks.
Widest single gap
LiveCodeBench (Vals AI): Claude Haiku 4.5 41.2% against DeepSeek V4 Pro 0423 87.5%, 46.3 points apart on a 0 to 100 scale.
Output price
Claude Haiku 4.5 $5.00 against DeepSeek V4 Pro 0423 $1.9105 per million output tokens, 2.6x apart.
Context window
Claude Haiku 4.5 takes 200K tokens, DeepSeek V4 Pro 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 Haiku 4.5BarsDeepSeek V4 Pro 0423Ties
Coding0121 tie
Intelligence02
Agentic work06
Reasoning06
Math03
Knowledge110
Web design08
General120
All shared boards2671 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index012
Coding benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
SWE-bench Verified66.6%Not measuredOnly Claude Haiku 4.5 measured
SWE-bench Multilingual64.7%Not measuredOnly Claude Haiku 4.5 measured
LiveBench coding48.1%Not measuredOnly Claude Haiku 4.5 measured
LMArena WebDev13301447DeepSeek V4 Pro 0423
AA Coding Index43.959.4DeepSeek V4 Pro 0423
LiveCodeBench (Vals AI)41.2%87.5%DeepSeek V4 Pro 0423
SWE-bench Verified (Vals AI)66.6%77.4%DeepSeek V4 Pro 0423
Terminal-Bench 4 (Vals AI)Not measured11.1%Only DeepSeek V4 Pro 0423 measured
Vibe Code Bench (Vals AI)11.4%49.9%DeepSeek V4 Pro 0423
Code migration (Vals AI)7.5%26.2%DeepSeek V4 Pro 0423
ProgramBench (Vals AI)0.0%0.0%Tie
Terminal-Bench 2.1 (Vals AI)43.8%50.2%DeepSeek V4 Pro 0423
SkillsBench (Vals AI)Not measured51.3%Only DeepSeek V4 Pro 0423 measured
Terminal-Bench35.5%Not measuredOnly Claude Haiku 4.5 measured
FrontierCodeNot measured17.6%Only DeepSeek V4 Pro 0423 measured
WeirdML45.4%48.9%DeepSeek V4 Pro 0423
ALE-BenchNot measured1006.1Only DeepSeek V4 Pro 0423 measured
LMArena Text: coding14541470DeepSeek V4 Pro 0423
LMArena Text: software and IT14391469DeepSeek V4 Pro 0423
LMArena WebDev: React13261441DeepSeek V4 Pro 0423
LMArena WebDev: HTML13181435DeepSeek V4 Pro 0423
Neither model has a result on the other 17 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
AA Intelligence Index16.930.4DeepSeek V4 Pro 0423
Epoch Capabilities Index142.2149.1DeepSeek V4 Pro 0423
Agentic workLike-for-likePublic boards, shown as published, kept out of the index06
Agentic work benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
AA Agentic Index8.026.3DeepSeek V4 Pro 0423
Design Arena: agents building web appsNot measured1000Only DeepSeek V4 Pro 0423 measured
Design Arena: agents building full-stack appsNot measured948Only DeepSeek V4 Pro 0423 measured
Design Arena: agents building Godot gamesNot measured1059Only DeepSeek V4 Pro 0423 measured
LiveBench research42.2%Not measuredOnly Claude Haiku 4.5 measured
LiveBench browser40.0%Not measuredOnly Claude Haiku 4.5 measured
LiveBench long agentic11.1%Not measuredOnly Claude Haiku 4.5 measured
Finance Agent (Vals AI)31.0%44.1%DeepSeek V4 Pro 0423
Tax Agent Bench (Vals AI)5.9%25.2%DeepSeek V4 Pro 0423
Legal Research Bench (Vals AI)10.6%23.1%DeepSeek V4 Pro 0423
Excel Modeling Benchmark (Vals AI)22.7%51.6%DeepSeek V4 Pro 0423
BALROG31.2%Not measuredOnly Claude Haiku 4.5 measured
DeepResearch Bench45.5%Not measuredOnly Claude Haiku 4.5 measured
Vending-Bench 2$459$3,285DeepSeek V4 Pro 0423
Neither model has a result on the other 25 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index06
Reasoning benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
GPQA Diamond72.2%89.4%DeepSeek V4 Pro 0423
LiveBench reasoning61.7%Not measuredOnly Claude Haiku 4.5 measured
MMMU Pro (Vals AI)46.1%Not measuredOnly Claude Haiku 4.5 measured
GPQA Diamond (Epoch AI run)71.2%90.9%DeepSeek V4 Pro 0423
ARC-AGI47.7%Not measuredOnly Claude Haiku 4.5 measured
ARC-AGI-24.0%Not measuredOnly Claude Haiku 4.5 measured
Chess puzzles8.0%19.0%DeepSeek V4 Pro 0423
Mystery game puzzlesNot measured17.0%Only DeepSeek V4 Pro 0423 measured
LMArena Text: hard prompts14211461DeepSeek V4 Pro 0423
LMArena Text: expert14431463DeepSeek V4 Pro 0423
LMArena Text: hard prompts in English14351466DeepSeek V4 Pro 0423
Neither model has a result on the other 10 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
AIME82.7%Not measuredOnly Claude Haiku 4.5 measured
MGSM (Vals AI)92.2%Not measuredOnly Claude Haiku 4.5 measured
OTIS Mock AIME66.7%96.7%DeepSeek V4 Pro 0423
FrontierMath, tiers 1 to 3Not measured45.3%Only DeepSeek V4 Pro 0423 measured
FrontierMath, tier 4Not measured2.4%Only DeepSeek V4 Pro 0423 measured
ProofBenchNot measured16.0%Only DeepSeek V4 Pro 0423 measured
MATH level 596.4%Not measuredOnly Claude Haiku 4.5 measured
LMArena Text: math13981443DeepSeek V4 Pro 0423
LMArena Text: mathematical work14301455DeepSeek V4 Pro 0423
Neither model has a result on the other 4 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index110
Knowledge benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
MMLU Pro78.7%87.3%DeepSeek V4 Pro 0423
LegalBench (Vals AI)81.2%80.3%Claude Haiku 4.5
CaseLaw (Vals AI)56.5%59.4%DeepSeek V4 Pro 0423
CorpFin (Vals AI)60.6%61.4%DeepSeek V4 Pro 0423
TaxEval (Vals AI)67.5%72.1%DeepSeek V4 Pro 0423
MortgageTax (Vals AI)62.2%Not measuredOnly Claude Haiku 4.5 measured
MedQA (Vals AI)79.6%Not measuredOnly Claude Haiku 4.5 measured
MedCode (Vals AI)32.7%40.5%DeepSeek V4 Pro 0423
SimpleQA Verified13.2%47.0%DeepSeek V4 Pro 0423
LMArena Text: science14061470DeepSeek V4 Pro 0423
LMArena Text: medicine and healthcare13941466DeepSeek V4 Pro 0423
LMArena Text: legal and government14001464DeepSeek V4 Pro 0423
LMArena Text: business and finance14001443DeepSeek V4 Pro 0423
Neither model has a result on the other 4 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index08
Web design benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
Design Arena: websites11221235DeepSeek V4 Pro 0423
Design Arena: UI components11001221DeepSeek V4 Pro 0423
Design Arena: game dev10971229DeepSeek V4 Pro 0423
Design Arena: data viz11241203DeepSeek V4 Pro 0423
Design Arena: 3D10721254DeepSeek V4 Pro 0423
Design Arena: all code categories11151240DeepSeek V4 Pro 0423
Design Arena: SVG10251139DeepSeek V4 Pro 0423
Design Arena: ASCII art11391148DeepSeek V4 Pro 0423
LiveBench design59.9%Not measuredOnly Claude Haiku 4.5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index120
General benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 0423Leader
LiveBench data analysis59.3%Not measuredOnly Claude Haiku 4.5 measured
Public Benefits Bench (Vals AI)54.3%62.9%DeepSeek V4 Pro 0423
SAGE (Vals AI)31.8%Not measuredOnly Claude Haiku 4.5 measured
MedScribe (Vals AI)85.2%75.1%Claude Haiku 4.5
Vals Multimodal Index42.9%Not measuredOnly Claude Haiku 4.5 measured
DTBench73.6%90.7%DeepSeek V4 Pro 0423
LMCA30.9%41.2%DeepSeek V4 Pro 0423
Surface Evolver BenchNot measured40.0%Only DeepSeek V4 Pro 0423 measured
ExploitBench13.7%Not measuredOnly Claude Haiku 4.5 measured
CL-bench LifeNot measured13.5%Only DeepSeek V4 Pro 0423 measured
LMArena Text13971451DeepSeek V4 Pro 0423
LMArena Text: instruction following14141445DeepSeek V4 Pro 0423
LMArena Text: creative writing13711446DeepSeek V4 Pro 0423
LMArena Text: English14141459DeepSeek V4 Pro 0423
LMArena Text: non-English13771439DeepSeek V4 Pro 0423
LMArena Text: Chinese14181481DeepSeek V4 Pro 0423
LMArena Text: German13741458DeepSeek V4 Pro 0423
LMArena Text: Japanese13381446DeepSeek V4 Pro 0423
LMArena Text: Korean13461409DeepSeek V4 Pro 0423
LMArena Text: French14111473DeepSeek V4 Pro 0423
LMArena Text: Spanish14211457DeepSeek V4 Pro 0423
LMArena Text: Russian13811453DeepSeek V4 Pro 0423
LMArena Text: Polish13431461DeepSeek V4 Pro 0423
LMArena Text: multi-turn14091468DeepSeek V4 Pro 0423
LMArena Text: longer queries14271458DeepSeek V4 Pro 0423
LMArena Text: writing and language13861447DeepSeek V4 Pro 0423
LMArena Text: entertainment and media13691431DeepSeek V4 Pro 0423
LMArena Document1420Not measuredOnly Claude Haiku 4.5 measured
Neither model has a result on the other 18 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Haiku 4.5 and DeepSeek V4 Pro 0423
BenchmarkClaude Haiku 4.5DeepSeek V4 Pro 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 Haiku 4.5$0.0035Fits in one request
DeepSeek V4 Pro 0423$0.0019Fits in one request

DeepSeek V4 Pro 0423 costs less

Repository review

50K input + 3K output tokens

Claude Haiku 4.5$0.065Fits in one request
DeepSeek V4 Pro 0423$0.053Fits in one request

DeepSeek V4 Pro 0423 costs less

Agent loop

200K input + 10K output tokens

Claude Haiku 4.5$0.250Exceeds the context window
DeepSeek V4 Pro 0423$0.210Fits in one request

DeepSeek V4 Pro 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 Haiku 4.5 and DeepSeek V4 Pro 0423
SpecificationAnthropic Claude Haiku 4.5DeepSeek V4 Pro 0423
MakerAnthropicDeepSeek
WeightsNo public weights listedOpen weights
ReleasedOctober 15, 2025April 24, 2026
Context window200K1.05M
Input $/Mtok$1.00$0.9553
Output $/Mtok$5.00$1.9105
Coding index49.835.3
Independent measurements behind it31
Cost per solved task$0.56$0.57
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Haiku 4.5 better than DeepSeek V4 Pro 0423 for coding?

Pick Claude Haiku 4.5 for coding: it scores 49.8 on the Vibe Coding Index against 35.3 for DeepSeek V4 Pro 0423. Pick DeepSeek V4 Pro 0423 when output tokens cost 2.6x less, or when the context window is larger, 1.05M against 200K tokens.

Which is cheaper, Claude Haiku 4.5 or DeepSeek V4 Pro 0423?

Claude Haiku 4.5 costs $1.00 per million input tokens and $5.00 per million output tokens; DeepSeek V4 Pro 0423 costs $0.9553 and $1.9105. Per solved SWE-bench Verified task, Claude Haiku 4.5 costs $0.56 and DeepSeek V4 Pro 0423 $0.57.

Which has the bigger context window, Claude Haiku 4.5 or DeepSeek V4 Pro 0423?

DeepSeek V4 Pro 0423, with 1.05M tokens against 200K.

How many benchmarks is the Claude Haiku 4.5 vs DeepSeek V4 Pro 0423 comparison based on?

70 benchmarks have a result for both models as of October 21, 2026. Claude Haiku 4.5's coding index rests on 3 independent benchmarks and DeepSeek V4 Pro 0423's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next