Enter opens the highlighted result. Esc closes.

Claude Haiku 5.5 vs Kimi K3

Claude Haiku 5.5 leads Kimi K3 on coding, 70.7 to 62.3 on the Vibe Coding Index, and Claude Haiku 5.5 costs 30.0x less per output token ($0.50 vs $15.00 per million). Updated October 22, 2026; 58 benchmarks have a result for both.

Replace Claude Haiku 5.5 Replace Kimi K3
Claude Haiku 5.5Anthropic
Coding
70.7
Overall
67.6
Output $/M
$0.50
29
Kimi K3Moonshot AI
Coding
62.3
Overall
66.5
Output $/M
$15.00
29
vs boards won of 58 shared

Claude Haiku 5.5 or Kimi K3: which to pick

Pick Claude Haiku 5.5 for coding: it scores 70.7 on the Vibe Coding Index against 62.3 for Kimi K3, though on 5 benchmarks against 6.

Pick Kimi K3 for a larger context window (1.05M against 1M tokens) and work measured by LMArena WebDev or IOI (Vals AI), where it leads.

Model AClaude Haiku 5.5Anthropic70.7/1005 independent benchmarksNo cost per solved taskModel BKimi K3Moonshot AI62.3/1006 independent benchmarks$0.81 per solved task
Shared results
58
Claude Haiku 5.5 only
5
Kimi K3 only
59
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 Haiku 5.5 70.7 from 5 benchmarks, Kimi K3 62.3 from 6 benchmarks.
Widest category
Knowledge, 14.5 points apart on average over 2 shared benchmarks.
Widest single gap
LiveBench browser: Claude Haiku 5.5 80.0% against Kimi K3 16.0%, 64.0 points apart on a 0 to 100 scale.
Output price
Claude Haiku 5.5 $0.50 against Kimi K3 $15.00 per million output tokens, 30.0x apart.
Context window
Claude Haiku 5.5 takes 1M tokens, Kimi K3 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 5.5BarsKimi K3Ties
Coding96
Intelligence11
Agentic work66
Reasoning34
Math40
Knowledge11
Web design16
General45
All shared boards2929
CodingLike-for-likeSix of these boards feed the Vibe Coding Index96
Coding benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
LiveBench coding91.4%59.3%Claude Haiku 5.5
LMArena WebDev15861655Kimi K3
AA Coding IndexNot measured76.2Only Kimi K3 measured
LiveCodeBench (Vals AI)Not measured87.2%Only Kimi K3 measured
SWE-bench Verified (Vals AI)Not measured93.4%Only Kimi K3 measured
Terminal-Bench 4 (Vals AI)36.4%17.2%Claude Haiku 5.5
Vibe Code Bench (Vals AI)90.4%85.0%Claude Haiku 5.5
Code migration (Vals AI)44.6%16.1%Claude Haiku 5.5
ProgramBench (Vals AI)2.5%2.0%Claude Haiku 5.5
IOI (Vals AI)47.3%48.9%Kimi K3
Terminal-Bench 2.1 (Vals AI)Not measured80.9%Only Kimi K3 measured
SWE-bench Multimodal (Vals AI)44.0%44.2%Kimi K3
Vibe Code Bench 1-100 (Vals AI)20.4%18.2%Claude Haiku 5.5
CyberBench (Vals AI)75.5%75.2%Claude Haiku 5.5
SciCode55.0%59.5%Kimi K3
FrontierCode46.4%44.2%Claude Haiku 5.5
FrontierSWE43.8%25.9%Claude Haiku 5.5
WeirdMLNot measured82.6%Only Kimi K3 measured
ALE-BenchNot measured1524.5Only Kimi K3 measured
LMArena Text: codingNot measured1508Only Kimi K3 measured
LMArena Text: software and ITNot measured1499Only Kimi K3 measured
LMArena WebDev: React16051670Kimi K3
LMArena WebDev: HTML15691644Kimi K3
LMArena WebDev: from an imageNot measured1579Only Kimi K3 measured
SciCode9.2%Not measuredOnly Claude Haiku 5.5 measured
Neither model has a result on the other 14 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
Intelligence benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
AA Intelligence Index43.443.6Kimi K3
Epoch Capabilities Index159.3157.9Claude Haiku 5.5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index66
Agentic work benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
AA Agentic IndexNot measured50.0Only Kimi K3 measured
Design Arena: agents building web appsNot measured1295Only Kimi K3 measured
Design Arena: agents building full-stack appsNot measured1315Only Kimi K3 measured
Design Arena: agents building mobile appsNot measured1269Only Kimi K3 measured
Design Arena: agents building Android appsNot measured1253Only Kimi K3 measured
Design Arena: agents building Godot gamesNot measured1223Only Kimi K3 measured
Design Arena: agents building gamesNot measured1252Only Kimi K3 measured
Design Arena: HTML slidesNot measured1206Only Kimi K3 measured
Design Arena: PowerPoint slidesNot measured1247Only Kimi K3 measured
Design Arena: spreadsheetsNot measured1070Only Kimi K3 measured
LiveBench research81.9%79.0%Claude Haiku 5.5
LiveBench browser80.0%16.0%Claude Haiku 5.5
LiveBench long agentic57.6%42.8%Claude Haiku 5.5
SREBench (Vals AI)5.7%Not measuredOnly Claude Haiku 5.5 measured
Finance Agent (Vals AI)54.1%53.1%Claude Haiku 5.5
Tax Agent Bench (Vals AI)31.9%33.4%Kimi K3
Legal Research Bench (Vals AI)43.3%46.1%Kimi K3
Harvey's Legal Agent Benchmark (Vals AI)4.2%12.9%Kimi K3
Excel Modeling Benchmark (Vals AI)59.0%66.7%Kimi K3
BioMysteryBench (Vals AI)70.4%72.6%Kimi K3
Terminal-Bench-Science (Vals AI)10.0%1.4%Claude Haiku 5.5
Vals Science Index38.5%Not measuredOnly Claude Haiku 5.5 measured
Time Horizon Index (Vals AI)Not measured10.5%Only Kimi K3 measured
APEX-Agents56.1%47.8%Claude Haiku 5.5
PostTrainBenchNot measured32.0%Only Kimi K3 measured
Vending-Bench 2$3,602$5,165Kimi K3
Neither model has a result on the other 15 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index34
Reasoning benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
GPQA DiamondNot measured92.9%Only Kimi K3 measured
LiveBench reasoning82.4%90.7%Kimi K3
MysteryMechanism (Vals AI)35.1%Not measuredOnly Claude Haiku 5.5 measured
MMMU Pro (Vals AI)Not measured88.2%Only Kimi K3 measured
GPQA Diamond (Artificial Analysis run)Not measured93.5%Only Kimi K3 measured
Humanity's Last Exam (Artificial Analysis run)44.4%46.9%Kimi K3
GPQA Diamond (Epoch AI run)89.6%93.1%Kimi K3
ARC-AGI89.0%94.5%Kimi K3
ARC-AGI-261.3%60.4%Claude Haiku 5.5
SimpleBenchNot measured60.7%Only Kimi K3 measured
Chess puzzles55.0%37.0%Claude Haiku 5.5
Mystery game puzzles30.0%26.0%Claude Haiku 5.5
EBR-bench43.8%Not measuredOnly Claude Haiku 5.5 measured
LMArena Text: hard promptsNot measured1495Only Kimi K3 measured
LMArena Text: expertNot measured1520Only Kimi K3 measured
LMArena Text: hard prompts in EnglishNot measured1496Only Kimi K3 measured
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index40
Math benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
OTIS Mock AIME98.9%97.2%Claude Haiku 5.5
FrontierMath, tiers 1 to 375.1%72.2%Claude Haiku 5.5
FrontierMath, tier 446.3%39.0%Claude Haiku 5.5
ProofBench89.0%87.0%Claude Haiku 5.5
LMArena Text: mathNot measured1494Only Kimi K3 measured
LMArena Text: mathematical workNot measured1505Only Kimi K3 measured
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index11
Knowledge benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
MMLU ProNot measured88.0%Only Kimi K3 measured
LegalBench (Vals AI)Not measured86.2%Only Kimi K3 measured
CorpFin (Vals AI)Not measured71.6%Only Kimi K3 measured
TaxEval (Vals AI)Not measured75.7%Only Kimi K3 measured
MortgageTax (Vals AI)Not measured66.3%Only Kimi K3 measured
MedCode (Vals AI)51.5%49.4%Claude Haiku 5.5
SimpleQA Verified23.8%50.6%Kimi K3
LMArena Text: scienceNot measured1508Only Kimi K3 measured
LMArena Text: medicine and healthcareNot measured1489Only Kimi K3 measured
LMArena Text: legal and governmentNot measured1503Only Kimi K3 measured
LMArena Text: business and financeNot measured1470Only Kimi K3 measured
Neither model has a result on the other 7 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index16
Web design benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
Design Arena: websites12621334Kimi K3
Design Arena: UI components12911341Kimi K3
Design Arena: game dev13131358Kimi K3
Design Arena: data viz12941351Kimi K3
Design Arena: 3D13421387Kimi K3
Design Arena: all code categories12781360Kimi K3
Design Arena: SVGNot measured1302Only Kimi K3 measured
LiveBench design83.1%77.2%Claude Haiku 5.5
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index45
General benchmarks for Claude Haiku 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
Vals Index58.7%53.0%Claude Haiku 5.5
LiveBench data analysis53.1%78.7%Kimi K3
Public Benefits Bench (Vals AI)63.2%68.2%Kimi K3
SAGE (Vals AI)45.8%52.8%Kimi K3
MedScribe (Vals AI)89.5%88.0%Claude Haiku 5.5
Vals Multimodal IndexNot measured73.4%Only Kimi K3 measured
Long context reasoning (AA-LCR)82.7%88.7%Kimi K3
DTBench95.0%91.2%Claude Haiku 5.5
LMCA51.1%52.7%Kimi K3
Surface Evolver BenchNot measured95.0%Only Kimi K3 measured
Furniture assembly47.5%34.2%Claude Haiku 5.5
LMArena TextNot measured1476Only Kimi K3 measured
LMArena Text: instruction followingNot measured1485Only Kimi K3 measured
LMArena Text: creative writingNot measured1457Only Kimi K3 measured
LMArena Text: EnglishNot measured1481Only Kimi K3 measured
LMArena Text: non-EnglishNot measured1466Only Kimi K3 measured
LMArena Text: ChineseNot measured1529Only Kimi K3 measured
LMArena Text: GermanNot measured1490Only Kimi K3 measured
LMArena Text: JapaneseNot measured1494Only Kimi K3 measured
LMArena Text: KoreanNot measured1460Only Kimi K3 measured
LMArena Text: FrenchNot measured1492Only Kimi K3 measured
LMArena Text: SpanishNot measured1472Only Kimi K3 measured
LMArena Text: RussianNot measured1485Only Kimi K3 measured
LMArena Text: PolishNot measured1458Only Kimi K3 measured
LMArena Text: multi-turnNot measured1489Only Kimi K3 measured
LMArena Text: longer queriesNot measured1494Only Kimi K3 measured
LMArena Text: writing and languageNot measured1471Only Kimi K3 measured
LMArena Text: entertainment and mediaNot measured1453Only Kimi K3 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 5.5 and Kimi K3
BenchmarkClaude Haiku 5.5Kimi K3Leader
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 5.5$0.0003Fits in one request
Kimi K3$0.011Fits in one request

Claude Haiku 5.5 costs less

Repository review

50K input + 3K output tokens

Claude Haiku 5.5$0.0065Fits in one request
Kimi K3$0.195Fits in one request

Claude Haiku 5.5 costs less

Agent loop

200K input + 10K output tokens

Claude Haiku 5.5$0.025Fits in one request
Kimi K3$0.750Fits in one request

Claude Haiku 5.5 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 5.5 and Kimi K3
SpecificationAnthropic Claude Haiku 5.5MoonshotAI Kimi K3
MakerAnthropicMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedOctober 7, 2026July 16, 2026
Context window1M1.05M
Input $/Mtok$0.10$3.00
Output $/Mtok$0.50$15.00
Coding index70.762.3
Independent measurements behind it56
Cost per solved taskNot computable$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Haiku 5.5 better than Kimi K3 for coding?

Pick Claude Haiku 5.5 for coding: it scores 70.7 on the Vibe Coding Index against 62.3 for Kimi K3, though on 5 benchmarks against 6. Pick Kimi K3 for a larger context window (1.05M against 1M tokens) and work measured by LMArena WebDev or IOI (Vals AI), where it leads.

Which is cheaper, Claude Haiku 5.5 or Kimi K3?

Claude Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens; Kimi K3 costs $3.00 and $15.00. Cost per solved task is known for Kimi K3 only.

Which has the bigger context window, Claude Haiku 5.5 or Kimi K3?

Kimi K3, with 1.05M tokens against 1M.

How many benchmarks is the Claude Haiku 5.5 vs Kimi K3 comparison based on?

58 benchmarks have a result for both models as of October 22, 2026. Claude Haiku 5.5's coding index rests on 5 independent benchmarks and Kimi K3's on 6. 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-5-5-vs-kimi-k3" title="Claude Haiku 5.5 vs Kimi K3, 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-5-5-vs-kimi-k3">LLMBenchmark.ai, Claude Haiku 5.5 vs Kimi K3</a></p>