Claude Haiku 5.5 vs Kimi K2.6
Claude Haiku 5.5 leads Kimi K2.6 on coding, 70.7 to 52.1 on the Vibe Coding Index, and Claude Haiku 5.5 costs 8.0x less per output token ($0.50 vs $4.00 per million). Updated October 22, 2026; 38 benchmarks have a result for both.
Claude Haiku 5.5 or Kimi K2.6: which to pick
Pick Claude Haiku 5.5 for coding: it scores 70.7 on the Vibe Coding Index against 52.1 for Kimi K2.6.
Pick Kimi K2.6 for work measured by Vending-Bench 2 or GPQA Diamond (Epoch AI run), where it leads.
- Shared results
- 38
- Claude Haiku 5.5 only
- 25
- Kimi K2.6 only
- 71
- 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.
Coding
Vibe Coding Index, six public coding benchmarks
Claude Haiku 5.5
Claude Haiku 5.5 scores 70.7 against 52.1.
5 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
Claude Haiku 5.5
Claude Haiku 5.5 leads, 43.4 against 27.0.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Kimi K2.6 has a AA Agentic Index result, 20.5.
Directional only
Reasoning
GPQA Diamond
No call
Only Kimi K2.6 has a GPQA Diamond result, 89.1%.
Directional only
Knowledge
MMLU Pro
No call
Only Kimi K2.6 has a MMLU Pro result, 87.6%.
Directional only
Web design
Design Arena: websites
Kimi K2.6
Kimi K2.6 leads, 1268 against 1262.
Same board, both measured
General
Vals Index
No call
Only Claude Haiku 5.5 has a Vals Index result, 58.7%.
Directional only
Long documents
Listed context window
Claude Haiku 5.5
Claude Haiku 5.5 takes 1M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Claude Haiku 5.5
Claude Haiku 5.5 $0.0003, Kimi K2.6 $0.0029 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Kimi K2.6 has the number, $0.35.
Directional only
At a glance
- Coding index
- Claude Haiku 5.5 70.7 from 5 benchmarks, Kimi K2.6 52.1 from 3 benchmarks.
- Widest category
- Math, 28.6 points apart on average over 4 shared benchmarks.
- Widest single gap
- ProofBench: Claude Haiku 5.5 89.0% against Kimi K2.6 16.0%, 73.0 points apart on a 0 to 100 scale.
- Output price
- Claude Haiku 5.5 $0.50 against Kimi K2.6 $4.00 per million output tokens, 8.0x apart.
- Context window
- Claude Haiku 5.5 takes 1M tokens, Kimi K2.6 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.
| Category | Claude Haiku 5.5 | Bars | Kimi K2.6 | Ties |
|---|---|---|---|---|
| Coding | 7 | 0 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 4 | 1 | ||
| Reasoning | 4 | 1 | ||
| Math | 4 | 0 | ||
| Knowledge | 1 | 1 | ||
| Web design | 5 | 1 | ||
| General | 6 | 1 | ||
| All shared boards | 33 | 5 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index70
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| LiveBench coding | 91.4% | Not measured | Only Claude Haiku 5.5 measured |
| LMArena WebDev | 1586 | 1509 | Claude Haiku 5.5 |
| AA Coding Index | Not measured | 61.8 | Only Kimi K2.6 measured |
| LiveCodeBench (Vals AI) | Not measured | 86.8% | Only Kimi K2.6 measured |
| SWE-bench Verified (Vals AI) | Not measured | 76.2% | Only Kimi K2.6 measured |
| Terminal-Bench 4 (Vals AI) | 36.4% | Not measured | Only Claude Haiku 5.5 measured |
| Vibe Code Bench (Vals AI) | 90.4% | 37.9% | Claude Haiku 5.5 |
| Code migration (Vals AI) | 44.6% | 27.8% | Claude Haiku 5.5 |
| ProgramBench (Vals AI) | 2.5% | 0.0% | Claude Haiku 5.5 |
| IOI (Vals AI) | 47.3% | Not measured | Only Claude Haiku 5.5 measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 53.6% | Only Kimi K2.6 measured |
| SWE-bench Multimodal (Vals AI) | 44.0% | Not measured | Only Claude Haiku 5.5 measured |
| Vibe Code Bench 1-100 (Vals AI) | 20.4% | Not measured | Only Claude Haiku 5.5 measured |
| CyberBench (Vals AI) | 75.5% | Not measured | Only Claude Haiku 5.5 measured |
| SciCode | 55.0% | 51.5% | Claude Haiku 5.5 |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 43.9% | Only Kimi K2.6 measured |
| FrontierCode | 46.4% | Not measured | Only Claude Haiku 5.5 measured |
| FrontierSWE | 43.8% | Not measured | Only Claude Haiku 5.5 measured |
| WeirdML | Not measured | 55.9% | Only Kimi K2.6 measured |
| ALE-Bench | Not measured | 1092.7 | Only Kimi K2.6 measured |
| LMArena Text: coding | Not measured | 1489 | Only Kimi K2.6 measured |
| LMArena Text: software and IT | Not measured | 1484 | Only Kimi K2.6 measured |
| LMArena WebDev: React | 1605 | 1505 | Claude Haiku 5.5 |
| LMArena WebDev: HTML | 1569 | 1485 | Claude Haiku 5.5 |
| LMArena WebDev: from an image | Not measured | 1521 | Only Kimi K2.6 measured |
| SciCode | 9.2% | Not measured | Only Claude Haiku 5.5 measured |
| Neither model has a result on the other 13 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| AA Intelligence Index | 43.4 | 27.0 | Claude Haiku 5.5 |
| Epoch Capabilities Index | 159.3 | 150.9 | Claude Haiku 5.5 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index41
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 20.5 | Only Kimi K2.6 measured |
| Design Arena: agents building web apps | Not measured | 1268 | Only Kimi K2.6 measured |
| Design Arena: agents building full-stack apps | Not measured | 1155 | Only Kimi K2.6 measured |
| Design Arena: agents building mobile apps | Not measured | 1156 | Only Kimi K2.6 measured |
| Design Arena: agents building Android apps | Not measured | 1161 | Only Kimi K2.6 measured |
| Design Arena: agents building Godot games | Not measured | 1120 | Only Kimi K2.6 measured |
| Design Arena: agents building games | Not measured | 1141 | Only Kimi K2.6 measured |
| Design Arena: HTML slides | Not measured | 1135 | Only Kimi K2.6 measured |
| Design Arena: PowerPoint slides | Not measured | 1103 | Only Kimi K2.6 measured |
| Design Arena: spreadsheets | Not measured | 1276 | Only Kimi K2.6 measured |
| LiveBench research | 81.9% | Not measured | Only Claude Haiku 5.5 measured |
| LiveBench browser | 80.0% | Not measured | Only Claude Haiku 5.5 measured |
| LiveBench long agentic | 57.6% | Not measured | Only Claude Haiku 5.5 measured |
| SREBench (Vals AI) | 5.7% | Not measured | Only Claude Haiku 5.5 measured |
| Finance Agent (Vals AI) | 54.1% | 44.9% | Claude Haiku 5.5 |
| Tax Agent Bench (Vals AI) | 31.9% | 12.4% | Claude Haiku 5.5 |
| Legal Research Bench (Vals AI) | 43.3% | 15.9% | Claude Haiku 5.5 |
| Harvey's Legal Agent Benchmark (Vals AI) | 4.2% | Not measured | Only Claude Haiku 5.5 measured |
| Excel Modeling Benchmark (Vals AI) | 59.0% | 57.9% | Claude Haiku 5.5 |
| BioMysteryBench (Vals AI) | 70.4% | Not measured | Only Claude Haiku 5.5 measured |
| Terminal-Bench-Science (Vals AI) | 10.0% | Not measured | Only Claude Haiku 5.5 measured |
| Vals Science Index | 38.5% | Not measured | Only Claude Haiku 5.5 measured |
| τ²-Bench Telecom | Not measured | 95.9% | Only Kimi K2.6 measured |
| APEX-Agents | 56.1% | Not measured | Only Claude Haiku 5.5 measured |
| OSWorld 2.0 | Not measured | 4.6% | Only Kimi K2.6 measured |
| Vending-Bench 2 | $3,602 | $6,205 | Kimi K2.6 |
| Neither model has a result on the other 15 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index41
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 89.1% | Only Kimi K2.6 measured |
| LiveBench reasoning | 82.4% | Not measured | Only Claude Haiku 5.5 measured |
| MysteryMechanism (Vals AI) | 35.1% | Not measured | Only Claude Haiku 5.5 measured |
| MMMU Pro (Vals AI) | Not measured | 86.3% | Only Kimi K2.6 measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 91.1% | Only Kimi K2.6 measured |
| Humanity's Last Exam (Artificial Analysis run) | 44.4% | 37.5% | Claude Haiku 5.5 |
| GPQA Diamond (Epoch AI run) | 89.6% | 90.8% | Kimi K2.6 |
| ARC-AGI | 89.0% | Not measured | Only Claude Haiku 5.5 measured |
| ARC-AGI-2 | 61.3% | Not measured | Only Claude Haiku 5.5 measured |
| Chess puzzles | 55.0% | 20.0% | Claude Haiku 5.5 |
| Mystery game puzzles | 30.0% | 18.0% | Claude Haiku 5.5 |
| EBR-bench | 43.8% | 2.4% | Claude Haiku 5.5 |
| LMArena Text: hard prompts | Not measured | 1470 | Only Kimi K2.6 measured |
| LMArena Text: expert | Not measured | 1491 | Only Kimi K2.6 measured |
| LMArena Text: hard prompts in English | Not measured | 1471 | Only Kimi K2.6 measured |
| LMArena Vision: diagrams | Not measured | 1298 | Only Kimi K2.6 measured |
| LMArena Vision: homework | Not measured | 1310 | Only Kimi K2.6 measured |
| Neither model has a result on the other 5 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| OTIS Mock AIME | 98.9% | 96.1% | Claude Haiku 5.5 |
| FrontierMath, tiers 1 to 3 | 75.1% | 57.2% | Claude Haiku 5.5 |
| FrontierMath, tier 4 | 46.3% | 25.6% | Claude Haiku 5.5 |
| ProofBench | 89.0% | 16.0% | Claude Haiku 5.5 |
| LMArena Text: math | Not measured | 1474 | Only Kimi K2.6 measured |
| LMArena Text: mathematical work | Not measured | 1480 | Only Kimi K2.6 measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 87.6% | Only Kimi K2.6 measured |
| LegalBench (Vals AI) | Not measured | 84.7% | Only Kimi K2.6 measured |
| CaseLaw (Vals AI) | Not measured | 61.2% | Only Kimi K2.6 measured |
| CorpFin (Vals AI) | Not measured | 66.7% | Only Kimi K2.6 measured |
| TaxEval (Vals AI) | Not measured | 74.7% | Only Kimi K2.6 measured |
| MortgageTax (Vals AI) | Not measured | 65.8% | Only Kimi K2.6 measured |
| MedCode (Vals AI) | 51.5% | 40.1% | Claude Haiku 5.5 |
| SimpleQA Verified | 23.8% | 34.9% | Kimi K2.6 |
| LMArena Text: science | Not measured | 1468 | Only Kimi K2.6 measured |
| LMArena Text: medicine and healthcare | Not measured | 1457 | Only Kimi K2.6 measured |
| LMArena Text: legal and government | Not measured | 1462 | Only Kimi K2.6 measured |
| LMArena Text: business and finance | Not measured | 1445 | Only Kimi K2.6 measured |
| Neither model has a result on the other 6 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index51
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| Design Arena: websites | 1262 | 1268 | Kimi K2.6 |
| Design Arena: UI components | 1291 | 1263 | Claude Haiku 5.5 |
| Design Arena: game dev | 1313 | 1250 | Claude Haiku 5.5 |
| Design Arena: data viz | 1294 | 1249 | Claude Haiku 5.5 |
| Design Arena: 3D | 1342 | 1267 | Claude Haiku 5.5 |
| Design Arena: all code categories | 1278 | 1271 | Claude Haiku 5.5 |
| Design Arena: SVG | Not measured | 1181 | Only Kimi K2.6 measured |
| Design Arena: ASCII art | Not measured | 1165 | Only Kimi K2.6 measured |
| LiveBench design | 83.1% | Not measured | Only Claude Haiku 5.5 measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index61
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| Vals Index | 58.7% | Not measured | Only Claude Haiku 5.5 measured |
| LiveBench data analysis | 53.1% | Not measured | Only Claude Haiku 5.5 measured |
| Public Benefits Bench (Vals AI) | 63.2% | 56.6% | Claude Haiku 5.5 |
| SAGE (Vals AI) | 45.8% | 50.2% | Kimi K2.6 |
| MedScribe (Vals AI) | 89.5% | 78.2% | Claude Haiku 5.5 |
| Vals Multimodal Index | Not measured | 56.4% | Only Kimi K2.6 measured |
| IFBench | Not measured | 76.0% | Only Kimi K2.6 measured |
| Long context reasoning (AA-LCR) | 82.7% | 81.0% | Claude Haiku 5.5 |
| DTBench | 95.0% | 90.9% | Claude Haiku 5.5 |
| LMCA | 51.1% | 37.3% | Claude Haiku 5.5 |
| Furniture assembly | 47.5% | 21.7% | Claude Haiku 5.5 |
| ExploitBench | Not measured | 18.4% | Only Kimi K2.6 measured |
| LMArena Text | Not measured | 1455 | Only Kimi K2.6 measured |
| LMArena Text: instruction following | Not measured | 1451 | Only Kimi K2.6 measured |
| LMArena Text: creative writing | Not measured | 1434 | Only Kimi K2.6 measured |
| LMArena Vision | Not measured | 1282 | Only Kimi K2.6 measured |
| LMArena Text: English | Not measured | 1460 | Only Kimi K2.6 measured |
| LMArena Text: non-English | Not measured | 1446 | Only Kimi K2.6 measured |
| LMArena Text: Chinese | Not measured | 1520 | Only Kimi K2.6 measured |
| LMArena Text: German | Not measured | 1451 | Only Kimi K2.6 measured |
| LMArena Text: Japanese | Not measured | 1443 | Only Kimi K2.6 measured |
| LMArena Text: Korean | Not measured | 1427 | Only Kimi K2.6 measured |
| LMArena Text: French | Not measured | 1471 | Only Kimi K2.6 measured |
| LMArena Text: Spanish | Not measured | 1464 | Only Kimi K2.6 measured |
| LMArena Text: Russian | Not measured | 1446 | Only Kimi K2.6 measured |
| LMArena Text: Polish | Not measured | 1438 | Only Kimi K2.6 measured |
| LMArena Text: multi-turn | Not measured | 1453 | Only Kimi K2.6 measured |
| LMArena Text: longer queries | Not measured | 1468 | Only Kimi K2.6 measured |
| LMArena Text: writing and language | Not measured | 1439 | Only Kimi K2.6 measured |
| LMArena Text: entertainment and media | Not measured | 1428 | Only Kimi K2.6 measured |
| LMArena Vision: English | Not measured | 1287 | Only Kimi K2.6 measured |
| LMArena Vision: Chinese | Not measured | 1346 | Only Kimi K2.6 measured |
| LMArena Vision: reading text | Not measured | 1291 | Only Kimi K2.6 measured |
| LMArena Vision: humor | Not measured | 1280 | Only Kimi K2.6 measured |
| LMArena Vision: creative writing | Not measured | 1273 | Only Kimi K2.6 measured |
| LMArena Document | Not measured | 1451 | Only Kimi K2.6 measured |
| Neither model has a result on the other 10 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Claude Haiku 5.5 | Kimi K2.6 | Leader |
|---|---|---|---|
| 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 costs less
Repository review
50K input + 3K output tokens
Claude Haiku 5.5 costs less
Agent loop
200K input + 10K output tokens
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.
| Specification | Anthropic Claude Haiku 5.5 | MoonshotAI Kimi K2.6 |
|---|---|---|
| Maker | Anthropic | Moonshot AI |
| Weights | No public weights listed | Open weights |
| Released | October 7, 2026 | April 20, 2026 |
| Context window | 1M | 262K |
| Input $/Mtok | $0.10 | $0.95 |
| Output $/Mtok | $0.50 | $4.00 |
| Coding index | 70.7 | 52.1 |
| Independent measurements behind it | 5 | 3 |
| Cost per solved task | Not computable | $0.35 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Claude Haiku 5.5 better than Kimi K2.6 for coding?
Pick Claude Haiku 5.5 for coding: it scores 70.7 on the Vibe Coding Index against 52.1 for Kimi K2.6. Pick Kimi K2.6 for work measured by Vending-Bench 2 or GPQA Diamond (Epoch AI run), where it leads.
Which is cheaper, Claude Haiku 5.5 or Kimi K2.6?
Claude Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens; Kimi K2.6 costs $0.95 and $4.00. Cost per solved task is known for Kimi K2.6 only.
Which has the bigger context window, Claude Haiku 5.5 or Kimi K2.6?
Claude Haiku 5.5, with 1M tokens against 262K.
How many benchmarks is the Claude Haiku 5.5 vs Kimi K2.6 comparison based on?
38 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 K2.6's on 3. Where only one model has a result, the page says so instead of naming a winner.