LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Haiku 4.5 vs DeepSeek V4.1 Flash

Updated October 21, 2026. 69 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 V4.1 Flash
Claude Haiku 4.5Anthropic
Coding
49.8
Overall
16.9
Output $/M
$5.00
0
DeepSeek V4.1 FlashDeepSeek
Coding
77.2
Overall
24.7
Output $/M
$0.525
69
vs boards won of 69 shared

Decision reading

Pick DeepSeek V4.1 Flash for coding: it scores 77.2 on the Vibe Coding Index against 49.8 for Claude Haiku 4.5, though on 2 benchmarks against 3.

Claude Haiku 4.5 is neither cheaper nor larger in context, and leads no benchmark both models were measured on.

Model AClaude Haiku 4.5Anthropic49.8/1003 independent benchmarks$0.56 per solved taskModel BDeepSeek V4.1 FlashDeepSeek77.2/1002 independent benchmarksNo cost per solved task
Shared results
69
Claude Haiku 4.5 only
25
DeepSeek V4.1 Flash only
32
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 V4.1 Flash 77.2 from 2 benchmarks.
Widest category
Coding, 33.6 points apart on average over 10 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): Claude Haiku 4.5 11.4% against DeepSeek V4.1 Flash 84.7%, 73.3 points apart on a 0 to 100 scale.
Output price
Claude Haiku 4.5 $5.00 against DeepSeek V4.1 Flash $0.525 per million output tokens, 9.5x apart.
Context window
Claude Haiku 4.5 takes 200K tokens, DeepSeek V4.1 Flash 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 V4.1 FlashTies
Coding010
Intelligence02
Agentic work07
Reasoning08
Math03
Knowledge07
Web design09
General023
All shared boards069
CodingLike-for-likeFour of these boards feed the Vibe Coding Index010
Coding benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
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%90.2%DeepSeek V4.1 Flash
LMArena WebDev13301618DeepSeek V4.1 Flash
AA Coding Index43.9Not measuredOnly Claude Haiku 4.5 measured
LiveCodeBench (Vals AI)41.2%Not measuredOnly Claude Haiku 4.5 measured
SWE-bench Verified (Vals AI)66.6%Not measuredOnly Claude Haiku 4.5 measured
Terminal-Bench 4 (Vals AI)Not measured19.7%Only DeepSeek V4.1 Flash measured
Vibe Code Bench (Vals AI)11.4%84.7%DeepSeek V4.1 Flash
Code migration (Vals AI)7.5%45.6%DeepSeek V4.1 Flash
ProgramBench (Vals AI)0.0%0.5%DeepSeek V4.1 Flash
IOI (Vals AI)Not measured40.3%Only DeepSeek V4.1 Flash measured
Terminal-Bench 2.1 (Vals AI)43.8%74.5%DeepSeek V4.1 Flash
SWE-bench Multimodal (Vals AI)Not measured46.5%Only DeepSeek V4.1 Flash measured
Vibe Code Bench 1-100 (Vals AI)Not measured16.4%Only DeepSeek V4.1 Flash measured
SkillsBench (Vals AI)Not measured69.8%Only DeepSeek V4.1 Flash measured
CyberBench (Vals AI)Not measured73.7%Only DeepSeek V4.1 Flash measured
SciCodeNot measured51.9%Only DeepSeek V4.1 Flash measured
Terminal-Bench35.5%Not measuredOnly Claude Haiku 4.5 measured
FrontierCodeNot measured43.9%Only DeepSeek V4.1 Flash measured
WeirdML45.4%Not measuredOnly Claude Haiku 4.5 measured
ALE-BenchNot measured1092.3Only DeepSeek V4.1 Flash measured
LMArena Text: coding14541502DeepSeek V4.1 Flash
LMArena Text: software and IT14391496DeepSeek V4.1 Flash
LMArena WebDev: React13261628DeepSeek V4.1 Flash
LMArena WebDev: HTML13181612DeepSeek V4.1 Flash
LMArena WebDev: from an imageNot measured1606Only DeepSeek V4.1 Flash 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 Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
AA Intelligence Index16.924.7DeepSeek V4.1 Flash
Epoch Capabilities Index142.2155.9DeepSeek V4.1 Flash
Agentic workLike-for-likePublic boards, shown as published, kept out of the index07
Agentic work benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
AA Agentic Index8.0Not measuredOnly Claude Haiku 4.5 measured
LiveBench research42.2%86.8%DeepSeek V4.1 Flash
LiveBench browser40.0%84.0%DeepSeek V4.1 Flash
LiveBench long agentic11.1%40.1%DeepSeek V4.1 Flash
SREBench (Vals AI)Not measured0.8%Only DeepSeek V4.1 Flash measured
Finance Agent (Vals AI)31.0%53.5%DeepSeek V4.1 Flash
Tax Agent Bench (Vals AI)5.9%26.6%DeepSeek V4.1 Flash
Legal Research Bench (Vals AI)10.6%41.4%DeepSeek V4.1 Flash
Excel Modeling Benchmark (Vals AI)22.7%57.2%DeepSeek V4.1 Flash
BioMysteryBench (Vals AI)Not measured67.8%Only DeepSeek V4.1 Flash measured
Terminal-Bench-Science (Vals AI)Not measured0.0%Only DeepSeek V4.1 Flash measured
Vals Science IndexNot measured29.6%Only DeepSeek V4.1 Flash measured
APEX-AgentsNot measured39.5%Only DeepSeek V4.1 Flash measured
BALROG31.2%Not measuredOnly Claude Haiku 4.5 measured
DeepResearch Bench45.5%Not measuredOnly Claude Haiku 4.5 measured
Vending-Bench 2$459Not measuredOnly Claude Haiku 4.5 measured
Neither model has a result on the other 23 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index08
Reasoning benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
GPQA Diamond72.2%Not measuredOnly Claude Haiku 4.5 measured
LiveBench reasoning61.7%86.7%DeepSeek V4.1 Flash
MysteryMechanism (Vals AI)Not measured21.2%Only DeepSeek V4.1 Flash measured
MMMU Pro (Vals AI)46.1%Not measuredOnly Claude Haiku 4.5 measured
Humanity's Last Exam (Artificial Analysis run)Not measured39.2%Only DeepSeek V4.1 Flash measured
GPQA Diamond (Epoch AI run)71.2%89.8%DeepSeek V4.1 Flash
ARC-AGI47.7%94.5%DeepSeek V4.1 Flash
ARC-AGI-24.0%72.9%DeepSeek V4.1 Flash
Chess puzzles8.0%48.0%DeepSeek V4.1 Flash
Mystery game puzzlesNot measured43.0%Only DeepSeek V4.1 Flash measured
LMArena Text: hard prompts14211483DeepSeek V4.1 Flash
LMArena Text: expert14431501DeepSeek V4.1 Flash
LMArena Text: hard prompts in English14351488DeepSeek V4.1 Flash
LMArena Vision: diagramsNot measured1297Only DeepSeek V4.1 Flash measured
LMArena Vision: homeworkNot measured1307Only DeepSeek V4.1 Flash measured
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
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%98.3%DeepSeek V4.1 Flash
FrontierMath, tiers 1 to 3Not measured67.4%Only DeepSeek V4.1 Flash measured
FrontierMath, tier 4Not measured26.8%Only DeepSeek V4.1 Flash measured
ProofBenchNot measured54.0%Only DeepSeek V4.1 Flash measured
MATH level 596.4%Not measuredOnly Claude Haiku 4.5 measured
LMArena Text: math13981478DeepSeek V4.1 Flash
LMArena Text: mathematical work14301485DeepSeek V4.1 Flash
Neither model has a result on the other 4 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index07
Knowledge benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
MMLU Pro78.7%Not measuredOnly Claude Haiku 4.5 measured
LegalBench (Vals AI)81.2%83.3%DeepSeek V4.1 Flash
CaseLaw (Vals AI)56.5%Not measuredOnly Claude Haiku 4.5 measured
CorpFin (Vals AI)60.6%Not measuredOnly Claude Haiku 4.5 measured
TaxEval (Vals AI)67.5%Not measuredOnly Claude Haiku 4.5 measured
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%41.2%DeepSeek V4.1 Flash
SimpleQA Verified13.2%42.9%DeepSeek V4.1 Flash
LMArena Text: science14061469DeepSeek V4.1 Flash
LMArena Text: medicine and healthcare13941484DeepSeek V4.1 Flash
LMArena Text: legal and government14001443DeepSeek V4.1 Flash
LMArena Text: business and finance14001461DeepSeek V4.1 Flash
Neither model has a result on the other 4 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index09
Web design benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
Design Arena: websites11221309DeepSeek V4.1 Flash
Design Arena: UI components11001314DeepSeek V4.1 Flash
Design Arena: game dev10971331DeepSeek V4.1 Flash
Design Arena: data viz11241290DeepSeek V4.1 Flash
Design Arena: 3D10721356DeepSeek V4.1 Flash
Design Arena: all code categories11151320DeepSeek V4.1 Flash
Design Arena: SVG10251243DeepSeek V4.1 Flash
Design Arena: ASCII art11391265DeepSeek V4.1 Flash
LiveBench design59.9%79.2%DeepSeek V4.1 Flash
GeneralLike-for-likePublic boards, shown as published, kept out of the index023
General benchmarks for Claude Haiku 4.5 and DeepSeek V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
LiveBench data analysis59.3%79.3%DeepSeek V4.1 Flash
Public Benefits Bench (Vals AI)54.3%64.3%DeepSeek V4.1 Flash
SAGE (Vals AI)31.8%47.9%DeepSeek V4.1 Flash
MedScribe (Vals AI)85.2%85.5%DeepSeek V4.1 Flash
Vals Multimodal Index42.9%Not measuredOnly Claude Haiku 4.5 measured
Long context reasoning (AA-LCR)Not measured84.0%Only DeepSeek V4.1 Flash measured
DTBench73.6%89.9%DeepSeek V4.1 Flash
LMCA30.9%47.0%DeepSeek V4.1 Flash
Surface Evolver BenchNot measured46.3%Only DeepSeek V4.1 Flash measured
Furniture assemblyNot measured34.2%Only DeepSeek V4.1 Flash measured
ExploitBench13.7%Not measuredOnly Claude Haiku 4.5 measured
LMArena Text13971464DeepSeek V4.1 Flash
LMArena Text: instruction following14141473DeepSeek V4.1 Flash
LMArena Text: creative writing13711437DeepSeek V4.1 Flash
LMArena VisionNot measured1278Only DeepSeek V4.1 Flash measured
LMArena Text: English14141473DeepSeek V4.1 Flash
LMArena Text: non-English13771449DeepSeek V4.1 Flash
LMArena Text: Chinese14181501DeepSeek V4.1 Flash
LMArena Text: German13741470DeepSeek V4.1 Flash
LMArena Text: Japanese13381432DeepSeek V4.1 Flash
LMArena Text: Korean13461436DeepSeek V4.1 Flash
LMArena Text: French14111465DeepSeek V4.1 Flash
LMArena Text: Spanish14211459DeepSeek V4.1 Flash
LMArena Text: Russian13811467DeepSeek V4.1 Flash
LMArena Text: Polish13431457DeepSeek V4.1 Flash
LMArena Text: multi-turn14091461DeepSeek V4.1 Flash
LMArena Text: longer queries14271473DeepSeek V4.1 Flash
LMArena Text: writing and language13861450DeepSeek V4.1 Flash
LMArena Text: entertainment and media13691445DeepSeek V4.1 Flash
LMArena Vision: EnglishNot measured1288Only DeepSeek V4.1 Flash measured
LMArena Vision: ChineseNot measured1327Only DeepSeek V4.1 Flash measured
LMArena Vision: reading textNot measured1282Only DeepSeek V4.1 Flash measured
LMArena Vision: humorNot measured1260Only DeepSeek V4.1 Flash measured
LMArena Vision: creative writingNot measured1275Only DeepSeek V4.1 Flash measured
LMArena Document1420Not measuredOnly Claude Haiku 4.5 measured
Neither model has a result on the other 11 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 V4.1 Flash
BenchmarkClaude Haiku 4.5DeepSeek V4.1 FlashLeader
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 V4.1 Flash$0.0003Fits in one request

DeepSeek V4.1 Flash costs less

Repository review

50K input + 3K output tokens

Claude Haiku 4.5$0.065Fits in one request
DeepSeek V4.1 Flash$0.0040Fits in one request

DeepSeek V4.1 Flash costs less

Agent loop

200K input + 10K output tokens

Claude Haiku 4.5$0.250Exceeds the context window
DeepSeek V4.1 Flash$0.015Fits in one request

DeepSeek V4.1 Flash 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 V4.1 Flash
SpecificationAnthropic Claude Haiku 4.5DeepSeek V4.1 Flash
MakerAnthropicDeepSeek
WeightsNo public weights listedOpen weights
ReleasedOctober 15, 2025September 10, 2026
Context window200K1.05M
Input $/Mtok$1.00$0.049
Output $/Mtok$5.00$0.525
Coding index49.877.2
Independent measurements behind it32
Cost per solved task$0.56Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Haiku 4.5 better than DeepSeek V4.1 Flash for coding?

Pick DeepSeek V4.1 Flash for coding: it scores 77.2 on the Vibe Coding Index against 49.8 for Claude Haiku 4.5, though on 2 benchmarks against 3. Claude Haiku 4.5 is neither cheaper nor larger in context, and leads no benchmark both models were measured on.

Which is cheaper, Claude Haiku 4.5 or DeepSeek V4.1 Flash?

Claude Haiku 4.5 costs $1.00 per million input tokens and $5.00 per million output tokens; DeepSeek V4.1 Flash costs $0.049 and $0.525. Cost per solved task is known for Claude Haiku 4.5 only.

Which has the bigger context window, Claude Haiku 4.5 or DeepSeek V4.1 Flash?

DeepSeek V4.1 Flash, with 1.05M tokens against 200K.

How many benchmarks is the Claude Haiku 4.5 vs DeepSeek V4.1 Flash comparison based on?

69 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 V4.1 Flash's on 2. 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-haiku-4-5-vs-deepseek-v4-1-flash" title="Claude Haiku 4.5 vs DeepSeek V4.1 Flash, 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-haiku-4-5-vs-deepseek-v4-1-flash">LLMBenchmark.ai, Claude Haiku 4.5 vs DeepSeek V4.1 Flash</a></p>