LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Haiku 4.5 vs Claude Sonnet 4.6

Updated October 21, 2026. 87 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 Claude Sonnet 4.6
Claude Haiku 4.5Anthropic
Coding
49.8
Overall
16.9
Output $/M
$5.00
0
Claude Sonnet 4.6Anthropic
Coding
57.8
Overall
30.1
Output $/M
$15.00
87
vs boards won of 87 shared

Decision reading

Pick Claude Sonnet 4.6 for coding: it scores 57.8 on the Vibe Coding Index against 49.8 for Claude Haiku 4.5, though on 2 benchmarks against 3.

Pick Claude Haiku 4.5 when output tokens cost 3.0x less, or when a solved task costs $0.56 against $1.68.

Model AClaude Haiku 4.5Anthropic49.8/1003 independent benchmarks$0.56 per solved taskModel BClaude Sonnet 4.6Anthropic57.8/1002 independent benchmarks$1.68 per solved task
Shared results
87
Claude Haiku 4.5 only
7
Claude Sonnet 4.6 only
30
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, Claude Sonnet 4.6 57.8 from 2 benchmarks.
Widest category
Agentic work, 24.2 points apart on average over 10 shared benchmarks.
Widest single gap
ARC-AGI-2: Claude Haiku 4.5 4.0% against Claude Sonnet 4.6 60.4%, 56.4 points apart on a 0 to 100 scale.
Output price
Claude Haiku 4.5 $5.00 against Claude Sonnet 4.6 $15.00 per million output tokens, 3.0x apart.
Context window
Claude Haiku 4.5 takes 200K tokens, Claude Sonnet 4.6 1M.

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 4.5BarsClaude Sonnet 4.6Ties
Coding015
Intelligence02
Agentic work010
Reasoning010
Math04
Knowledge012
Web design09
General025
All shared boards087
CodingLike-for-likeFour of these boards feed the Vibe Coding Index015
Coding benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
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%66.6%Claude Sonnet 4.6
LMArena WebDev13301523Claude Sonnet 4.6
AA Coding Index43.963.0Claude Sonnet 4.6
LiveCodeBench (Vals AI)41.2%82.1%Claude Sonnet 4.6
SWE-bench Verified (Vals AI)66.6%77.4%Claude Sonnet 4.6
Vibe Code Bench (Vals AI)11.4%55.8%Claude Sonnet 4.6
Code migration (Vals AI)7.5%39.9%Claude Sonnet 4.6
ProgramBench (Vals AI)0.0%0.5%Claude Sonnet 4.6
Terminal-Bench 2.1 (Vals AI)43.8%57.3%Claude Sonnet 4.6
SkillsBench (Vals AI)Not measured49.0%Only Claude Sonnet 4.6 measured
SciCodeNot measured50.1%Only Claude Sonnet 4.6 measured
Terminal-Bench Hard (Artificial Analysis run)Not measured53.0%Only Claude Sonnet 4.6 measured
Terminal-Bench35.5%53.4%Claude Sonnet 4.6
FrontierCodeNot measured24.3%Only Claude Sonnet 4.6 measured
WeirdML45.4%66.1%Claude Sonnet 4.6
LMArena Text: coding14541503Claude Sonnet 4.6
LMArena Text: software and IT14391491Claude Sonnet 4.6
LMArena WebDev: React13261512Claude Sonnet 4.6
LMArena WebDev: HTML13181527Claude Sonnet 4.6
LMArena WebDev: from an imageNot measured1529Only Claude Sonnet 4.6 measured
Neither model has a result on the other 16 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
AA Intelligence Index16.930.1Claude Sonnet 4.6
Epoch Capabilities Index142.2152.8Claude Sonnet 4.6
Agentic workLike-for-likePublic boards, shown as published, kept out of the index010
Agentic work benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
AA Agentic Index8.031.8Claude Sonnet 4.6
Design Arena: agents building web appsNot measured1190Only Claude Sonnet 4.6 measured
Design Arena: agents building full-stack appsNot measured1213Only Claude Sonnet 4.6 measured
Design Arena: agents building mobile appsNot measured1212Only Claude Sonnet 4.6 measured
Design Arena: agents building Android appsNot measured1202Only Claude Sonnet 4.6 measured
Design Arena: agents building Godot gamesNot measured1226Only Claude Sonnet 4.6 measured
Design Arena: agents building gamesNot measured1177Only Claude Sonnet 4.6 measured
LiveBench research42.2%63.6%Claude Sonnet 4.6
LiveBench browser40.0%68.0%Claude Sonnet 4.6
LiveBench long agentic11.1%23.1%Claude Sonnet 4.6
Finance Agent (Vals AI)31.0%51.0%Claude Sonnet 4.6
Tax Agent Bench (Vals AI)5.9%25.8%Claude Sonnet 4.6
Legal Research Bench (Vals AI)10.6%38.5%Claude Sonnet 4.6
Excel Modeling Benchmark (Vals AI)22.7%60.2%Claude Sonnet 4.6
τ²-Bench TelecomNot measured79.5%Only Claude Sonnet 4.6 measured
APEX-AgentsNot measured43.0%Only Claude Sonnet 4.6 measured
OSWorld 2.0Not measured9.3%Only Claude Sonnet 4.6 measured
BALROG31.2%Not measuredOnly Claude Haiku 4.5 measured
DeepResearch Bench45.5%54.9%Claude Sonnet 4.6
OSWorldNot measured72.1%Only Claude Sonnet 4.6 measured
Vending-Bench 2$459$7,204Claude Sonnet 4.6
Neither model has a result on the other 18 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index010
Reasoning benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
GPQA Diamond72.2%85.6%Claude Sonnet 4.6
LiveBench reasoning61.7%86.4%Claude Sonnet 4.6
MMMU Pro (Vals AI)46.1%83.6%Claude Sonnet 4.6
GPQA Diamond (Artificial Analysis run)Not measured87.5%Only Claude Sonnet 4.6 measured
Humanity's Last Exam (Artificial Analysis run)Not measured33.6%Only Claude Sonnet 4.6 measured
GPQA Diamond (Epoch AI run)71.2%87.4%Claude Sonnet 4.6
ARC-AGI47.7%86.5%Claude Sonnet 4.6
ARC-AGI-24.0%60.4%Claude Sonnet 4.6
Chess puzzles8.0%23.0%Claude Sonnet 4.6
Mystery game puzzlesNot measured16.0%Only Claude Sonnet 4.6 measured
LMArena Text: hard prompts14211485Claude Sonnet 4.6
LMArena Text: expert14431500Claude Sonnet 4.6
LMArena Text: hard prompts in English14351493Claude Sonnet 4.6
LMArena Vision: diagramsNot measured1306Only Claude Sonnet 4.6 measured
LMArena Vision: homeworkNot measured1311Only Claude Sonnet 4.6 measured
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index04
Math benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
AIME82.7%92.3%Claude Sonnet 4.6
MGSM (Vals AI)92.2%Not measuredOnly Claude Haiku 4.5 measured
OTIS Mock AIME66.7%85.8%Claude Sonnet 4.6
ProofBenchNot measured45.0%Only Claude Sonnet 4.6 measured
MATH level 596.4%Not measuredOnly Claude Haiku 4.5 measured
LMArena Text: math13981462Claude Sonnet 4.6
LMArena Text: mathematical work14301483Claude Sonnet 4.6
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index012
Knowledge benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
MMLU Pro78.7%87.3%Claude Sonnet 4.6
LegalBench (Vals AI)81.2%82.1%Claude Sonnet 4.6
CaseLaw (Vals AI)56.5%64.0%Claude Sonnet 4.6
CorpFin (Vals AI)60.6%65.3%Claude Sonnet 4.6
TaxEval (Vals AI)67.5%77.1%Claude Sonnet 4.6
MortgageTax (Vals AI)62.2%67.7%Claude Sonnet 4.6
MedQA (Vals AI)79.6%92.1%Claude Sonnet 4.6
MedCode (Vals AI)32.7%Not measuredOnly Claude Haiku 4.5 measured
SimpleQA Verified13.2%35.5%Claude Sonnet 4.6
LMArena Text: science14061474Claude Sonnet 4.6
LMArena Text: medicine and healthcare13941451Claude Sonnet 4.6
LMArena Text: legal and government14001470Claude Sonnet 4.6
LMArena Text: business and finance14001464Claude Sonnet 4.6
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 Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
Design Arena: websites11221285Claude Sonnet 4.6
Design Arena: UI components11001271Claude Sonnet 4.6
Design Arena: game dev10971259Claude Sonnet 4.6
Design Arena: data viz11241278Claude Sonnet 4.6
Design Arena: 3D10721234Claude Sonnet 4.6
Design Arena: all code categories11151276Claude Sonnet 4.6
Design Arena: SVG10251196Claude Sonnet 4.6
Design Arena: ASCII art11391235Claude Sonnet 4.6
LiveBench design59.9%71.8%Claude Sonnet 4.6
GeneralLike-for-likePublic boards, shown as published, kept out of the index025
General benchmarks for Claude Haiku 4.5 and Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
LiveBench data analysis59.3%76.1%Claude Sonnet 4.6
Public Benefits Bench (Vals AI)54.3%62.5%Claude Sonnet 4.6
SAGE (Vals AI)31.8%46.6%Claude Sonnet 4.6
MedScribe (Vals AI)85.2%Not measuredOnly Claude Haiku 4.5 measured
Vals Multimodal Index42.9%60.6%Claude Sonnet 4.6
IFBenchNot measured56.6%Only Claude Sonnet 4.6 measured
Long context reasoning (AA-LCR)Not measured80.0%Only Claude Sonnet 4.6 measured
DTBench73.6%89.9%Claude Sonnet 4.6
LMCA30.9%46.5%Claude Sonnet 4.6
ExploitBench13.7%23.6%Claude Sonnet 4.6
LMArena Text13971458Claude Sonnet 4.6
LMArena Text: instruction following14141476Claude Sonnet 4.6
LMArena Text: creative writing13711435Claude Sonnet 4.6
LMArena VisionNot measured1283Only Claude Sonnet 4.6 measured
LMArena Text: English14141472Claude Sonnet 4.6
LMArena Text: non-English13771440Claude Sonnet 4.6
LMArena Text: Chinese14181491Claude Sonnet 4.6
LMArena Text: German13741428Claude Sonnet 4.6
LMArena Text: Japanese13381420Claude Sonnet 4.6
LMArena Text: Korean13461410Claude Sonnet 4.6
LMArena Text: French14111465Claude Sonnet 4.6
LMArena Text: Spanish14211464Claude Sonnet 4.6
LMArena Text: Russian13811440Claude Sonnet 4.6
LMArena Text: Polish13431443Claude Sonnet 4.6
LMArena Text: multi-turn14091464Claude Sonnet 4.6
LMArena Text: longer queries14271479Claude Sonnet 4.6
LMArena Text: writing and language13861444Claude Sonnet 4.6
LMArena Text: entertainment and media13691431Claude Sonnet 4.6
LMArena Vision: EnglishNot measured1298Only Claude Sonnet 4.6 measured
LMArena Vision: ChineseNot measured1307Only Claude Sonnet 4.6 measured
LMArena Vision: reading textNot measured1296Only Claude Sonnet 4.6 measured
LMArena Vision: humorNot measured1248Only Claude Sonnet 4.6 measured
LMArena Vision: recognizing thingsNot measured1195Only Claude Sonnet 4.6 measured
LMArena Vision: creative writingNot measured1272Only Claude Sonnet 4.6 measured
LMArena Document14201482Claude Sonnet 4.6
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 Claude Sonnet 4.6
BenchmarkClaude Haiku 4.5Claude Sonnet 4.6Leader
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
Claude Sonnet 4.6$0.011Fits in one request

Claude Haiku 4.5 costs less

Repository review

50K input + 3K output tokens

Claude Haiku 4.5$0.065Fits in one request
Claude Sonnet 4.6$0.195Fits in one request

Claude Haiku 4.5 costs less

Agent loop

200K input + 10K output tokens

Claude Haiku 4.5$0.250Exceeds the context window
Claude Sonnet 4.6$0.750Fits in one request

Claude Haiku 4.5 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 Claude Sonnet 4.6
SpecificationAnthropic Claude Haiku 4.5Anthropic Claude Sonnet 4.6
MakerAnthropicAnthropic
WeightsNo public weights listedNo public weights listed
ReleasedOctober 15, 2025February 17, 2026
Context window200K1M
Input $/Mtok$1.00$3.00
Output $/Mtok$5.00$15.00
Coding index49.857.8
Independent measurements behind it32
Cost per solved task$0.56$1.68
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Haiku 4.5 better than Claude Sonnet 4.6 for coding?

Pick Claude Sonnet 4.6 for coding: it scores 57.8 on the Vibe Coding Index against 49.8 for Claude Haiku 4.5, though on 2 benchmarks against 3. Pick Claude Haiku 4.5 when output tokens cost 3.0x less, or when a solved task costs $0.56 against $1.68.

Which is cheaper, Claude Haiku 4.5 or Claude Sonnet 4.6?

Claude Haiku 4.5 costs $1.00 per million input tokens and $5.00 per million output tokens; Claude Sonnet 4.6 costs $3.00 and $15.00. Per solved SWE-bench Verified task, Claude Haiku 4.5 costs $0.56 and Claude Sonnet 4.6 $1.68.

Which has the bigger context window, Claude Haiku 4.5 or Claude Sonnet 4.6?

Claude Sonnet 4.6, with 1M tokens against 200K.

How many benchmarks is the Claude Haiku 4.5 vs Claude Sonnet 4.6 comparison based on?

87 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 Claude Sonnet 4.6'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-claude-sonnet-4-6" title="Claude Haiku 4.5 vs Claude Sonnet 4.6, 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-claude-sonnet-4-6">LLMBenchmark.ai, Claude Haiku 4.5 vs Claude Sonnet 4.6</a></p>