LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Haiku 4.5 vs Kimi K2.5

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 Kimi K2.5
Claude Haiku 4.5Anthropic
Coding
49.8
Overall
16.9
Output $/M
$5.00
3
Kimi K2.5Moonshot AI
Coding
58.5
Overall
23.5
Output $/M
$2.50
67
vs boards won of 70 shared

Decision reading

Pick Kimi K2.5 for coding: it scores 58.5 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 what matters is Code migration (Vals AI) or Terminal-Bench 2.1 (Vals AI), where it leads.

Model AClaude Haiku 4.5Anthropic49.8/1003 independent benchmarks$0.56 per solved taskModel BKimi K2.5Moonshot AI58.5/1002 independent benchmarks$0.24 per solved task
Shared results
70
Claude Haiku 4.5 only
24
Kimi K2.5 only
28
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, Kimi K2.5 58.5 from 2 benchmarks.
Widest category
Web design, 13.6 points apart on average over 8 shared benchmarks.
Widest single gap
LiveCodeBench (Vals AI): Claude Haiku 4.5 41.2% against Kimi K2.5 83.9%, 42.7 points apart on a 0 to 100 scale.
Output price
Claude Haiku 4.5 $5.00 against Kimi K2.5 $2.50 per million output tokens, 2.0x apart.
Context window
Claude Haiku 4.5 takes 200K tokens, Kimi K2.5 262K.

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 4.5BarsKimi K2.5Ties
Coding213
Intelligence02
Agentic work04
Reasoning07
Math03
Knowledge012
Web design08
General118
All shared boards367
CodingLike-for-likeFour of these boards feed the Vibe Coding Index213
Coding benchmarks for Claude Haiku 4.5 and Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
SWE-bench Verified66.6%70.8%Kimi K2.5
SWE-bench Multilingual64.7%67.3%Kimi K2.5
LiveBench coding48.1%Not measuredOnly Claude Haiku 4.5 measured
LMArena WebDev13301437Kimi K2.5
AA Coding Index43.946.8Kimi K2.5
LiveCodeBench (Vals AI)41.2%83.9%Kimi K2.5
SWE-bench Verified (Vals AI)66.6%70.0%Kimi K2.5
Vibe Code Bench (Vals AI)11.4%17.5%Kimi K2.5
Code migration (Vals AI)7.5%7.0%Claude Haiku 4.5
ProgramBench (Vals AI)0.0%Not measuredOnly Claude Haiku 4.5 measured
Terminal-Bench 2.1 (Vals AI)43.8%42.0%Claude Haiku 4.5
Terminal-Bench Hard (Artificial Analysis run)Not measured34.8%Only Kimi K2.5 measured
Terminal-Bench35.5%43.2%Kimi K2.5
WeirdML45.4%45.6%Kimi K2.5
ALE-BenchNot measured821.6Only Kimi K2.5 measured
LMArena Text: coding14541483Kimi K2.5
LMArena Text: software and IT14391471Kimi K2.5
LMArena WebDev: React13261429Kimi K2.5
LMArena WebDev: HTML13181426Kimi K2.5
LMArena WebDev: from an imageNot measured1439Only Kimi K2.5 measured
Neither model has a result on the other 18 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Claude Haiku 4.5 and Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
AA Intelligence Index16.923.5Kimi K2.5
Epoch Capabilities Index142.2148.1Kimi K2.5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index04
Agentic work benchmarks for Claude Haiku 4.5 and Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
AA Agentic Index8.0Not measuredOnly Claude Haiku 4.5 measured
Design Arena: agents building web appsNot measured1136Only Kimi K2.5 measured
Design Arena: agents building full-stack appsNot measured1114Only Kimi K2.5 measured
Design Arena: agents building mobile appsNot measured1147Only Kimi K2.5 measured
Design Arena: agents building Android appsNot measured1124Only Kimi K2.5 measured
Design Arena: agents building Godot gamesNot measured1219Only Kimi K2.5 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%35.8%Kimi K2.5
Tax Agent Bench (Vals AI)5.9%Not measuredOnly Claude Haiku 4.5 measured
Legal Research Bench (Vals AI)10.6%15.9%Kimi K2.5
Excel Modeling Benchmark (Vals AI)22.7%28.5%Kimi K2.5
τ²-Bench TelecomNot measured95.9%Only Kimi K2.5 measured
BALROG31.2%Not measuredOnly Claude Haiku 4.5 measured
DeepResearch Bench45.5%Not measuredOnly Claude Haiku 4.5 measured
OSWorldNot measured63.3%Only Kimi K2.5 measured
Vending-Bench 2$459$1,198Kimi K2.5
Neither model has a result on the other 21 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index07
Reasoning benchmarks for Claude Haiku 4.5 and Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
GPQA Diamond72.2%84.1%Kimi K2.5
LiveBench reasoning61.7%Not measuredOnly Claude Haiku 4.5 measured
MMMU Pro (Vals AI)46.1%84.3%Kimi K2.5
GPQA Diamond (Artificial Analysis run)Not measured87.9%Only Kimi K2.5 measured
Humanity's Last Exam (Artificial Analysis run)Not measured30.7%Only Kimi K2.5 measured
GPQA Diamond (Epoch AI run)71.2%Not measuredOnly Claude Haiku 4.5 measured
ARC-AGI47.7%65.3%Kimi K2.5
ARC-AGI-24.0%11.8%Kimi K2.5
SimpleBenchNot measured46.8%Only Kimi K2.5 measured
Chess puzzles8.0%Not measuredOnly Claude Haiku 4.5 measured
LMArena Text: hard prompts14211453Kimi K2.5
LMArena Text: expert14431466Kimi K2.5
LMArena Text: hard prompts in English14351460Kimi K2.5
LMArena Vision: diagramsNot measured1282Only Kimi K2.5 measured
LMArena Vision: homeworkNot measured1297Only Kimi K2.5 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 Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
AIME82.7%95.6%Kimi K2.5
MGSM (Vals AI)92.2%Not measuredOnly Claude Haiku 4.5 measured
OTIS Mock AIME66.7%Not measuredOnly Claude Haiku 4.5 measured
MATH level 596.4%Not measuredOnly Claude Haiku 4.5 measured
LMArena Text: math13981470Kimi K2.5
LMArena Text: mathematical work14301473Kimi K2.5
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index012
Knowledge benchmarks for Claude Haiku 4.5 and Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
MMLU Pro78.7%85.9%Kimi K2.5
LegalBench (Vals AI)81.2%Not measuredOnly Claude Haiku 4.5 measured
CaseLaw (Vals AI)56.5%58.7%Kimi K2.5
CorpFin (Vals AI)60.6%68.3%Kimi K2.5
TaxEval (Vals AI)67.5%74.2%Kimi K2.5
MortgageTax (Vals AI)62.2%66.5%Kimi K2.5
MedQA (Vals AI)79.6%94.4%Kimi K2.5
MedCode (Vals AI)32.7%39.3%Kimi K2.5
SimpleQA Verified13.2%34.3%Kimi K2.5
LMArena Text: science14061459Kimi K2.5
LMArena Text: medicine and healthcare13941448Kimi K2.5
LMArena Text: legal and government14001441Kimi K2.5
LMArena Text: business and finance14001427Kimi K2.5
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 Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
Design Arena: websites11221249Kimi K2.5
Design Arena: UI components11001236Kimi K2.5
Design Arena: game dev10971207Kimi K2.5
Design Arena: data viz11241219Kimi K2.5
Design Arena: 3D10721205Kimi K2.5
Design Arena: all code categories11151236Kimi K2.5
Design Arena: SVG10251149Kimi K2.5
Design Arena: ASCII art11391165Kimi K2.5
LiveBench design59.9%Not measuredOnly Claude Haiku 4.5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index118
General benchmarks for Claude Haiku 4.5 and Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
LiveBench data analysis59.3%Not measuredOnly Claude Haiku 4.5 measured
Public Benefits Bench (Vals AI)54.3%Not measuredOnly Claude Haiku 4.5 measured
SAGE (Vals AI)31.8%49.9%Kimi K2.5
MedScribe (Vals AI)85.2%76.4%Claude Haiku 4.5
Vals Multimodal Index42.9%Not measuredOnly Claude Haiku 4.5 measured
IFBenchNot measured70.2%Only Kimi K2.5 measured
Long context reasoning (AA-LCR)Not measured78.0%Only Kimi K2.5 measured
DTBench73.6%Not measuredOnly Claude Haiku 4.5 measured
LMCA30.9%Not measuredOnly Claude Haiku 4.5 measured
Fiction.LiveBenchNot measured86.1%Only Kimi K2.5 measured
CL-benchNot measured19.3%Only Kimi K2.5 measured
ExploitBench13.7%Not measuredOnly Claude Haiku 4.5 measured
CL-bench LifeNot measured13.2%Only Kimi K2.5 measured
LMArena Text13971445Kimi K2.5
LMArena Text: instruction following14141431Kimi K2.5
LMArena Text: creative writing13711422Kimi K2.5
LMArena VisionNot measured1269Only Kimi K2.5 measured
LMArena Text: English14141453Kimi K2.5
LMArena Text: non-English13771433Kimi K2.5
LMArena Text: Chinese14181494Kimi K2.5
LMArena Text: German13741441Kimi K2.5
LMArena Text: Japanese13381423Kimi K2.5
LMArena Text: Korean13461412Kimi K2.5
LMArena Text: French14111454Kimi K2.5
LMArena Text: Spanish14211450Kimi K2.5
LMArena Text: Russian13811435Kimi K2.5
LMArena Text: Polish13431445Kimi K2.5
LMArena Text: multi-turn14091444Kimi K2.5
LMArena Text: longer queries14271445Kimi K2.5
LMArena Text: writing and language13861425Kimi K2.5
LMArena Text: entertainment and media13691417Kimi K2.5
LMArena Vision: EnglishNot measured1274Only Kimi K2.5 measured
LMArena Vision: ChineseNot measured1328Only Kimi K2.5 measured
LMArena Vision: reading textNot measured1276Only Kimi K2.5 measured
LMArena Vision: humorNot measured1284Only Kimi K2.5 measured
LMArena Vision: captioningNot measured1257Only Kimi K2.5 measured
LMArena Vision: recognizing thingsNot measured1285Only Kimi K2.5 measured
LMArena Vision: creative writingNot measured1269Only Kimi K2.5 measured
LMArena Document1420Not measuredOnly Claude Haiku 4.5 measured
Neither model has a result on the other 7 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 Kimi K2.5
BenchmarkClaude Haiku 4.5Kimi K2.5Leader
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
Kimi K2.5$0.0018Fits in one request

Kimi K2.5 costs less

Repository review

50K input + 3K output tokens

Claude Haiku 4.5$0.065Fits in one request
Kimi K2.5$0.033Fits in one request

Kimi K2.5 costs less

Agent loop

200K input + 10K output tokens

Claude Haiku 4.5$0.250Exceeds the context window
Kimi K2.5$0.125Fits in one request

Kimi K2.5 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 Kimi K2.5
SpecificationAnthropic Claude Haiku 4.5MoonshotAI Kimi K2.5
MakerAnthropicMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedOctober 15, 2025January 27, 2026
Context window200K262K
Input $/Mtok$1.00$0.50
Output $/Mtok$5.00$2.50
Coding index49.858.5
Independent measurements behind it32
Cost per solved task$0.56$0.24
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Haiku 4.5 better than Kimi K2.5 for coding?

Pick Kimi K2.5 for coding: it scores 58.5 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 what matters is Code migration (Vals AI) or Terminal-Bench 2.1 (Vals AI), where it leads.

Which is cheaper, Claude Haiku 4.5 or Kimi K2.5?

Claude Haiku 4.5 costs $1.00 per million input tokens and $5.00 per million output tokens; Kimi K2.5 costs $0.50 and $2.50. Per solved SWE-bench Verified task, Claude Haiku 4.5 costs $0.56 and Kimi K2.5 $0.24.

Which has the bigger context window, Claude Haiku 4.5 or Kimi K2.5?

Kimi K2.5, with 262K tokens against 200K.

How many benchmarks is the Claude Haiku 4.5 vs Kimi K2.5 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 Kimi K2.5'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-kimi-k2-5" title="Claude Haiku 4.5 vs Kimi K2.5, 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-kimi-k2-5">LLMBenchmark.ai, Claude Haiku 4.5 vs Kimi K2.5</a></p>