GPT-5.2 vs Kimi K2.5
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.
Decision reading
GPT-5.2 and Kimi K2.5 are level for coding, 58.9 against 58.5 on the Vibe Coding Index, so price and context decide.
Pick Kimi K2.5 when output tokens cost 5.6x less, or when a solved task costs $0.24 against $1.91. Pick GPT-5.2 when the context window is larger, 400K against 262K tokens, or when what matters is SWE-bench Verified or LiveCodeBench (Vals AI), where it leads.
- Shared results
- 87
- GPT-5.2 only
- 24
- Kimi K2.5 only
- 11
- 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.
Coding
Vibe Coding Index, four public coding benchmarks
No call
Level: 58.9 against 58.5, inside 2 points.
2 benchmarks behind each score
Intelligence
AA Intelligence Index
GPT-5.2
GPT-5.2 leads, 30.4 against 23.5.
Same board, both measured
Reasoning
GPQA Diamond
GPT-5.2
GPT-5.2 leads, 91.7% against 84.1%.
Same board, both measured
Math
AIME
GPT-5.2
GPT-5.2 leads, 96.9% against 95.6%.
Same board, both measured
Knowledge
MMLU Pro
GPT-5.2
GPT-5.2 leads, 86.2% against 85.9%.
Same board, both measured
Web design
Design Arena: websites
Kimi K2.5
Kimi K2.5 leads, 1249 against 1196.
Same board, both measured
Long documents
Listed context window
GPT-5.2
GPT-5.2 takes 400K tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Kimi K2.5
GPT-5.2 $0.0088, Kimi K2.5 $0.0018 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
Kimi K2.5
GPT-5.2 $1.91, Kimi K2.5 $0.24.
Published attempt cost and pass rate
At a glance
- Coding index
- GPT-5.2 58.9 from 2 benchmarks, Kimi K2.5 58.5 from 2 benchmarks.
- Widest category
- Coding, 10.6 points apart on average over 13 shared benchmarks.
- Widest single gap
- ARC-AGI-2: GPT-5.2 52.9% against Kimi K2.5 11.8%, 41.1 points apart on a 0 to 100 scale.
- Output price
- GPT-5.2 $14.00 against Kimi K2.5 $2.50 per million output tokens, 5.6x apart.
- Context window
- GPT-5.2 takes 400K 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.
| Category | GPT-5.2 | Bars | Kimi K2.5 | Ties |
|---|---|---|---|---|
| Coding | 9 | 4 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 1 | 6 | ||
| Reasoning | 6 | 6 | ||
| Math | 1 | 2 | ||
| Knowledge | 6 | 6 | ||
| Web design | 0 | 8 | ||
| General | 6 | 24 | ||
| All shared boards | 31 | 56 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index94
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| SWE-bench Verified | 72.8% | 70.8% | GPT-5.2 |
| SWE-bench Multilingual | 66.7% | 67.3% | Kimi K2.5 |
| LMArena WebDev | 1417 | 1437 | Kimi K2.5 |
| AA Coding Index | Not measured | 46.8 | Only Kimi K2.5 measured |
| LiveCodeBench (Vals AI) | 85.4% | 83.9% | GPT-5.2 |
| SWE-bench Verified (Vals AI) | 75.8% | 70.0% | GPT-5.2 |
| Vibe Code Bench (Vals AI) | 53.5% | 17.5% | GPT-5.2 |
| Code migration (Vals AI) | Not measured | 7.0% | Only Kimi K2.5 measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 42.0% | Only Kimi K2.5 measured |
| LiveCodeBench (Artificial Analysis run) | 89.4% | Not measured | Only GPT-5.2 measured |
| Terminal-Bench Hard (Artificial Analysis run) | 47.0% | 34.8% | GPT-5.2 |
| Terminal-Bench | 64.9% | 43.2% | GPT-5.2 |
| GSO-Bench | 27.5% | Not measured | Only GPT-5.2 measured |
| WeirdML | 72.2% | 45.6% | GPT-5.2 |
| ALE-Bench | 1293.5 | 821.6 | GPT-5.2 |
| LMArena Text: coding | 1442 | 1483 | Kimi K2.5 |
| LMArena Text: software and IT | 1434 | 1471 | Kimi K2.5 |
| LMArena WebDev: React | Not measured | 1429 | Only Kimi K2.5 measured |
| LMArena WebDev: HTML | 1427 | 1426 | GPT-5.2 |
| LMArena WebDev: from an image | Not measured | 1439 | Only 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 index20
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| AA Intelligence Index | 30.4 | 23.5 | GPT-5.2 |
| Epoch Capabilities Index | 153.7 | 148.1 | GPT-5.2 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index16
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| Design Arena: agents building web apps | 1086 | 1136 | Kimi K2.5 |
| Design Arena: agents building full-stack apps | 1043 | 1114 | Kimi K2.5 |
| Design Arena: agents building mobile apps | 1094 | 1147 | Kimi K2.5 |
| Design Arena: agents building Android apps | 1064 | 1124 | Kimi K2.5 |
| Design Arena: agents building Godot games | 1142 | 1219 | Kimi K2.5 |
| Finance Agent (Vals AI) | Not measured | 35.8% | Only Kimi K2.5 measured |
| Legal Research Bench (Vals AI) | Not measured | 15.9% | Only Kimi K2.5 measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 28.5% | Only Kimi K2.5 measured |
| τ²-Bench Telecom | 84.8% | 95.9% | Kimi K2.5 |
| METR time horizons | 75.3% | Not measured | Only GPT-5.2 measured |
| Remote Labor Index | 2.5% | Not measured | Only GPT-5.2 measured |
| GDPval | 49.7% | Not measured | Only GPT-5.2 measured |
| DeepResearch Bench | 41.1% | Not measured | Only GPT-5.2 measured |
| OSWorld | Not measured | 63.3% | Only Kimi K2.5 measured |
| Vending-Bench 2 | $3,591 | $1,198 | GPT-5.2 |
| Neither model has a result on the other 24 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index66
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| GPQA Diamond | 91.7% | 84.1% | GPT-5.2 |
| MMMU Pro (Vals AI) | 86.7% | 84.3% | GPT-5.2 |
| GPQA Diamond (Artificial Analysis run) | 90.3% | 87.9% | GPT-5.2 |
| Humanity's Last Exam (Artificial Analysis run) | 37.7% | 30.7% | GPT-5.2 |
| GPQA Diamond (Epoch AI run) | 91.4% | Not measured | Only GPT-5.2 measured |
| ARC-AGI | 86.2% | 65.3% | GPT-5.2 |
| ARC-AGI-2 | 52.9% | 11.8% | GPT-5.2 |
| SimpleBench | 45.8% | 46.8% | Kimi K2.5 |
| Chess puzzles | 48.0% | Not measured | Only GPT-5.2 measured |
| Mystery game puzzles | 23.0% | Not measured | Only GPT-5.2 measured |
| EBR-bench | 23.0% | Not measured | Only GPT-5.2 measured |
| VPCT | 84.0% | Not measured | Only GPT-5.2 measured |
| LMArena Text: hard prompts | 1428 | 1453 | Kimi K2.5 |
| LMArena Text: expert | 1445 | 1466 | Kimi K2.5 |
| LMArena Text: hard prompts in English | 1431 | 1460 | Kimi K2.5 |
| LMArena Vision: diagrams | 1265 | 1282 | Kimi K2.5 |
| LMArena Vision: homework | 1276 | 1297 | Kimi K2.5 |
| Neither model has a result on the other 4 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index12
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| AIME | 96.9% | 95.6% | GPT-5.2 |
| MGSM (Vals AI) | 94.0% | Not measured | Only GPT-5.2 measured |
| AA Math Index | 99.0 | Not measured | Only GPT-5.2 measured |
| AIME 2025 (Artificial Analysis run) | 99.0% | Not measured | Only GPT-5.2 measured |
| OTIS Mock AIME | 96.1% | Not measured | Only GPT-5.2 measured |
| FrontierMath, tiers 1 to 3 | 67.4% | Not measured | Only GPT-5.2 measured |
| FrontierMath, tier 4 | 31.7% | Not measured | Only GPT-5.2 measured |
| ProofBench | 15.0% | Not measured | Only GPT-5.2 measured |
| LMArena Text: math | 1440 | 1470 | Kimi K2.5 |
| LMArena Text: mathematical work | 1442 | 1473 | Kimi K2.5 |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index66
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| MMLU Pro | 86.2% | 85.9% | GPT-5.2 |
| LegalBench (Vals AI) | 82.8% | Not measured | Only GPT-5.2 measured |
| CaseLaw (Vals AI) | 66.0% | 58.7% | GPT-5.2 |
| CorpFin (Vals AI) | 65.9% | 68.3% | Kimi K2.5 |
| TaxEval (Vals AI) | 75.8% | 74.2% | GPT-5.2 |
| MortgageTax (Vals AI) | 67.1% | 66.5% | GPT-5.2 |
| MedQA (Vals AI) | 94.1% | 94.4% | Kimi K2.5 |
| MedCode (Vals AI) | 49.8% | 39.3% | GPT-5.2 |
| MMLU Pro (Artificial Analysis run) | 87.4% | Not measured | Only GPT-5.2 measured |
| SimpleQA Verified | 37.1% | 34.3% | GPT-5.2 |
| LMArena Text: science | 1425 | 1459 | Kimi K2.5 |
| LMArena Text: medicine and healthcare | 1425 | 1448 | Kimi K2.5 |
| LMArena Text: legal and government | 1427 | 1441 | Kimi K2.5 |
| LMArena Text: business and finance | 1414 | 1427 | Kimi K2.5 |
| Neither model has a result on the other 3 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index08
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| Design Arena: websites | 1196 | 1249 | Kimi K2.5 |
| Design Arena: UI components | 1190 | 1236 | Kimi K2.5 |
| Design Arena: game dev | 1196 | 1207 | Kimi K2.5 |
| Design Arena: data viz | 1202 | 1219 | Kimi K2.5 |
| Design Arena: 3D | 1079 | 1205 | Kimi K2.5 |
| Design Arena: all code categories | 1171 | 1236 | Kimi K2.5 |
| Design Arena: SVG | 1137 | 1149 | Kimi K2.5 |
| Design Arena: ASCII art | 1153 | 1165 | Kimi K2.5 |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index624
| Benchmark | GPT-5.2 | Kimi K2.5 | Leader |
|---|---|---|---|
| SAGE (Vals AI) | 49.3% | 49.9% | Kimi K2.5 |
| MedScribe (Vals AI) | 84.4% | 76.4% | GPT-5.2 |
| IFBench | 75.4% | 70.2% | GPT-5.2 |
| Long context reasoning (AA-LCR) | 82.7% | 78.0% | GPT-5.2 |
| DTBench | 90.9% | Not measured | Only GPT-5.2 measured |
| LMCA | 43.9% | Not measured | Only GPT-5.2 measured |
| Fiction.LiveBench | Not measured | 86.1% | Only Kimi K2.5 measured |
| CL-bench | 18.2% | 19.3% | Kimi K2.5 |
| Furniture assembly | 38.3% | Not measured | Only GPT-5.2 measured |
| CL-bench Life | Not measured | 13.2% | Only Kimi K2.5 measured |
| LMArena Text | 1416 | 1445 | Kimi K2.5 |
| LMArena Text: instruction following | 1409 | 1431 | Kimi K2.5 |
| LMArena Text: creative writing | 1376 | 1422 | Kimi K2.5 |
| LMArena Vision | 1246 | 1269 | Kimi K2.5 |
| LMArena Text: English | 1422 | 1453 | Kimi K2.5 |
| LMArena Text: non-English | 1405 | 1433 | Kimi K2.5 |
| LMArena Text: Chinese | 1447 | 1494 | Kimi K2.5 |
| LMArena Text: German | 1444 | 1441 | GPT-5.2 |
| LMArena Text: Japanese | 1420 | 1423 | Kimi K2.5 |
| LMArena Text: Korean | 1374 | 1412 | Kimi K2.5 |
| LMArena Text: French | 1454 | 1454 | GPT-5.2 |
| LMArena Text: Spanish | 1413 | 1450 | Kimi K2.5 |
| LMArena Text: Russian | 1415 | 1435 | Kimi K2.5 |
| LMArena Text: Polish | 1426 | 1445 | Kimi K2.5 |
| LMArena Text: multi-turn | 1419 | 1444 | Kimi K2.5 |
| LMArena Text: longer queries | 1413 | 1445 | Kimi K2.5 |
| LMArena Text: writing and language | 1392 | 1425 | Kimi K2.5 |
| LMArena Text: entertainment and media | 1378 | 1417 | Kimi K2.5 |
| LMArena Vision: English | 1243 | 1274 | Kimi K2.5 |
| LMArena Vision: Chinese | 1305 | 1328 | Kimi K2.5 |
| LMArena Vision: reading text | 1257 | 1276 | Kimi K2.5 |
| LMArena Vision: humor | 1227 | 1284 | Kimi K2.5 |
| LMArena Vision: captioning | 1267 | 1257 | GPT-5.2 |
| LMArena Vision: recognizing things | 1232 | 1285 | Kimi K2.5 |
| LMArena Vision: creative writing | 1225 | 1269 | Kimi K2.5 |
| LMArena Document | 1405 | Not measured | Only GPT-5.2 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 | GPT-5.2 | Kimi K2.5 | 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
Kimi K2.5 costs less
Repository review
50K input + 3K output tokens
Kimi K2.5 costs less
Agent loop
200K input + 10K output tokens
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.
| Specification | OpenAI GPT-5.2 | MoonshotAI Kimi K2.5 |
|---|---|---|
| Maker | OpenAI | Moonshot AI |
| Weights | No public weights listed | Open weights |
| Released | December 10, 2025 | January 27, 2026 |
| Context window | 400K | 262K |
| Input $/Mtok | $1.75 | $0.50 |
| Output $/Mtok | $14.00 | $2.50 |
| Coding index | 58.9 | 58.5 |
| Independent measurements behind it | 2 | 2 |
| Cost per solved task | $1.91 | $0.24 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-5.2 better than Kimi K2.5 for coding?
GPT-5.2 and Kimi K2.5 are level for coding, 58.9 against 58.5 on the Vibe Coding Index, so price and context decide. Pick Kimi K2.5 when output tokens cost 5.6x less, or when a solved task costs $0.24 against $1.91. Pick GPT-5.2 when the context window is larger, 400K against 262K tokens, or when what matters is SWE-bench Verified or LiveCodeBench (Vals AI), where it leads.
Which is cheaper, GPT-5.2 or Kimi K2.5?
GPT-5.2 costs $1.75 per million input tokens and $14.00 per million output tokens; Kimi K2.5 costs $0.50 and $2.50. Per solved SWE-bench Verified task, GPT-5.2 costs $1.91 and Kimi K2.5 $0.24.
Which has the bigger context window, GPT-5.2 or Kimi K2.5?
GPT-5.2, with 400K tokens against 262K.
How many benchmarks is the GPT-5.2 vs Kimi K2.5 comparison based on?
87 benchmarks have a result for both models as of October 21, 2026. GPT-5.2's coding index rests on 2 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.