DeepSeek V4 Flash 0731 vs Kimi K2.5
DeepSeek V4 Flash 0731 leads Kimi K2.5 on coding, 51.8 to 38.6 on the Vibe Coding Index, and DeepSeek V4 Flash 0731 costs 10.2x less per output token ($0.28 vs $2.85 per million). Updated October 21, 2026; 34 benchmarks have a result for both.
DeepSeek V4 Flash 0731 or Kimi K2.5: which to pick
Pick DeepSeek V4 Flash 0731 for coding: it scores 51.8 on the Vibe Coding Index against 38.6 for Kimi K2.5.
Pick Kimi K2.5 for work measured by CorpFin (Vals AI) or TaxEval (Vals AI), where it leads.
- Shared results
- 34
- DeepSeek V4 Flash 0731 only
- 21
- Kimi K2.5 only
- 64
- Like-for-like categories
- 7 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
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 scores 51.8 against 38.6.
4 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 34.3 against 23.5.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only DeepSeek V4 Flash 0731 has a AA Agentic Index result, 41.0.
Directional only
Reasoning
GPQA Diamond
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 89.9% against 84.1%.
Same board, both measured
Math
AIME
No call
Only Kimi K2.5 has a AIME result, 95.6%.
Directional only
Knowledge
MMLU Pro
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 leads, 86.2% against 85.9%.
Same board, both measured
Web design
Design Arena: websites
Kimi K2.5
Kimi K2.5 leads, 1248 against 1238.
Same board, both measured
Long documents
Listed context window
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 takes 1.05M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 $0.0003, Kimi K2.5 $0.0020 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 $0.01, Kimi K2.5 $0.24.
Published attempt cost and pass rate
At a glance
- Coding index
- DeepSeek V4 Flash 0731 51.8 from 4 benchmarks, Kimi K2.5 38.6 from 3 benchmarks.
- Widest category
- Coding, 24.0 points apart on average over 8 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): DeepSeek V4 Flash 0731 74.7% against Kimi K2.5 17.5%, 57.2 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V4 Flash 0731 $0.28 against Kimi K2.5 $2.85 per million output tokens, 10.2x apart.
- Context window
- DeepSeek V4 Flash 0731 takes 1.05M 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 | DeepSeek V4 Flash 0731 | Bars | Kimi K2.5 | Ties |
|---|---|---|---|---|
| Coding | 8 | 0 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 3 | 0 | ||
| Reasoning | 6 | 0 | ||
| Knowledge | 2 | 3 | ||
| Web design | 3 | 5 | ||
| General | 2 | 0 | ||
| All shared boards | 26 | 8 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index80
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| SWE-bench Verified | Not measured | 70.8% | Only Kimi K2.5 measured |
| SWE-bench Multilingual | Not measured | 67.3% | Only Kimi K2.5 measured |
| LMArena WebDev | Not measured | 1437 | Only Kimi K2.5 measured |
| AA Coding Index | 69.1 | 46.8 | DeepSeek V4 Flash 0731 |
| LiveCodeBench (Vals AI) | 87.3% | 83.9% | DeepSeek V4 Flash 0731 |
| SWE-bench Verified (Vals AI) | 88.8% | 70.0% | DeepSeek V4 Flash 0731 |
| Terminal-Bench 4 (Vals AI) | 18.7% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Vibe Code Bench (Vals AI) | 74.7% | 17.5% | DeepSeek V4 Flash 0731 |
| Code migration (Vals AI) | 38.6% | 7.0% | DeepSeek V4 Flash 0731 |
| ProgramBench (Vals AI) | 0.0% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| IOI (Vals AI) | 32.7% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Terminal-Bench 2.1 (Vals AI) | 67.0% | 42.0% | DeepSeek V4 Flash 0731 |
| SkillsBench (Vals AI) | 50.7% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| CyberBench (Vals AI) | 72.1% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| SciCode | 50.3% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 34.8% | Only Kimi K2.5 measured |
| Terminal-Bench | Not measured | 43.2% | Only Kimi K2.5 measured |
| FrontierCode | 18.8% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| WeirdML | 63.0% | 45.6% | DeepSeek V4 Flash 0731 |
| ALE-Bench | 1306.1 | 821.6 | DeepSeek V4 Flash 0731 |
| LMArena Text: coding | Not measured | 1483 | Only Kimi K2.5 measured |
| LMArena Text: software and IT | Not measured | 1471 | Only Kimi K2.5 measured |
| LMArena WebDev: React | Not measured | 1429 | Only Kimi K2.5 measured |
| LMArena WebDev: HTML | Not measured | 1426 | Only Kimi K2.5 measured |
| LMArena WebDev: from an image | Not measured | 1439 | Only Kimi K2.5 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 | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| AA Intelligence Index | 34.3 | 23.5 | DeepSeek V4 Flash 0731 |
| Epoch Capabilities Index | 154.8 | 148.1 | DeepSeek V4 Flash 0731 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| AA Agentic Index | 41.0 | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Design Arena: agents building web apps | Not measured | 1136 | Only Kimi K2.5 measured |
| Design Arena: agents building full-stack apps | Not measured | 1114 | Only Kimi K2.5 measured |
| Design Arena: agents building mobile apps | Not measured | 1147 | Only Kimi K2.5 measured |
| Design Arena: agents building Android apps | Not measured | 1118 | Only Kimi K2.5 measured |
| Design Arena: agents building Godot games | Not measured | 1219 | Only Kimi K2.5 measured |
| Finance Agent (Vals AI) | 49.5% | 35.8% | DeepSeek V4 Flash 0731 |
| Tax Agent Bench (Vals AI) | 28.1% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Legal Research Bench (Vals AI) | 30.3% | 15.9% | DeepSeek V4 Flash 0731 |
| Excel Modeling Benchmark (Vals AI) | 57.0% | 28.5% | DeepSeek V4 Flash 0731 |
| BioMysteryBench (Vals AI) | 64.4% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Terminal-Bench-Science (Vals AI) | 4.3% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| τ²-Bench Telecom | Not measured | 95.9% | Only Kimi K2.5 measured |
| OSWorld | Not measured | 63.3% | Only Kimi K2.5 measured |
| Vending-Bench 2 | Not measured | $1,198 | Only Kimi K2.5 measured |
| Neither model has a result on the other 24 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index60
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| GPQA Diamond | 89.9% | 84.1% | DeepSeek V4 Flash 0731 |
| MMMU Pro (Vals AI) | Not measured | 84.3% | Only Kimi K2.5 measured |
| GPQA Diamond (Artificial Analysis run) | 90.8% | 87.9% | DeepSeek V4 Flash 0731 |
| Humanity's Last Exam (Artificial Analysis run) | 38.6% | 30.7% | DeepSeek V4 Flash 0731 |
| GPQA Diamond (Epoch AI run) | 91.0% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| ARC-AGI | 89.0% | 65.3% | DeepSeek V4 Flash 0731 |
| ARC-AGI-2 | 61.4% | 11.8% | DeepSeek V4 Flash 0731 |
| SimpleBench | 61.1% | 46.8% | DeepSeek V4 Flash 0731 |
| Chess puzzles | 32.0% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Mystery game puzzles | 34.0% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| LMArena Text: hard prompts | Not measured | 1453 | Only Kimi K2.5 measured |
| LMArena Text: expert | Not measured | 1466 | Only Kimi K2.5 measured |
| LMArena Text: hard prompts in English | Not measured | 1460 | Only Kimi K2.5 measured |
| LMArena Vision: diagrams | Not measured | 1282 | Only Kimi K2.5 measured |
| LMArena Vision: homework | Not measured | 1297 | Only Kimi K2.5 measured |
| Neither model has a result on the other 6 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| AIME | Not measured | 95.6% | Only Kimi K2.5 measured |
| OTIS Mock AIME | 94.4% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| FrontierMath, tiers 1 to 3 | 57.5% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| FrontierMath, tier 4 | 24.4% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| ProofBench | 56.0% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| LMArena Text: math | Not measured | 1470 | Only Kimi K2.5 measured |
| LMArena Text: mathematical work | Not measured | 1473 | Only Kimi K2.5 measured |
| Neither model has a result on the other 6 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index23
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| MMLU Pro | 86.2% | 85.9% | DeepSeek V4 Flash 0731 |
| LegalBench (Vals AI) | 77.7% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| CaseLaw (Vals AI) | Not measured | 58.7% | Only Kimi K2.5 measured |
| CorpFin (Vals AI) | 61.9% | 68.3% | Kimi K2.5 |
| TaxEval (Vals AI) | 70.7% | 74.2% | Kimi K2.5 |
| MortgageTax (Vals AI) | Not measured | 66.5% | Only Kimi K2.5 measured |
| MedQA (Vals AI) | Not measured | 94.4% | Only Kimi K2.5 measured |
| MedCode (Vals AI) | 41.4% | 39.3% | DeepSeek V4 Flash 0731 |
| SimpleQA Verified | 33.6% | 34.3% | Kimi K2.5 |
| LMArena Text: science | Not measured | 1459 | Only Kimi K2.5 measured |
| LMArena Text: medicine and healthcare | Not measured | 1448 | Only Kimi K2.5 measured |
| LMArena Text: legal and government | Not measured | 1441 | Only Kimi K2.5 measured |
| LMArena Text: business and finance | Not measured | 1427 | Only Kimi K2.5 measured |
| Neither model has a result on the other 4 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index35
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| Design Arena: websites | 1238 | 1248 | Kimi K2.5 |
| Design Arena: UI components | 1235 | 1236 | Kimi K2.5 |
| Design Arena: game dev | 1208 | 1207 | DeepSeek V4 Flash 0731 |
| Design Arena: data viz | 1180 | 1219 | Kimi K2.5 |
| Design Arena: 3D | 1209 | 1205 | DeepSeek V4 Flash 0731 |
| Design Arena: all code categories | 1230 | 1236 | Kimi K2.5 |
| Design Arena: SVG | 1185 | 1149 | DeepSeek V4 Flash 0731 |
| Design Arena: ASCII art | 1090 | 1166 | 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 index20
| Benchmark | DeepSeek V4 Flash 0731 | Kimi K2.5 | Leader |
|---|---|---|---|
| SAGE (Vals AI) | Not measured | 49.9% | Only Kimi K2.5 measured |
| MedScribe (Vals AI) | 80.4% | 76.4% | DeepSeek V4 Flash 0731 |
| IFBench | Not measured | 70.2% | Only Kimi K2.5 measured |
| Long context reasoning (AA-LCR) | 79.7% | 78.0% | DeepSeek V4 Flash 0731 |
| DTBench | 90.9% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| LMCA | 41.7% | Not measured | Only DeepSeek V4 Flash 0731 measured |
| Fiction.LiveBench | Not measured | 86.1% | Only Kimi K2.5 measured |
| CL-bench | Not measured | 19.3% | Only Kimi K2.5 measured |
| CL-bench Life | Not measured | 13.2% | Only Kimi K2.5 measured |
| LMArena Text | Not measured | 1445 | Only Kimi K2.5 measured |
| LMArena Text: instruction following | Not measured | 1431 | Only Kimi K2.5 measured |
| LMArena Text: creative writing | Not measured | 1422 | Only Kimi K2.5 measured |
| LMArena Vision | Not measured | 1269 | Only Kimi K2.5 measured |
| LMArena Text: English | Not measured | 1453 | Only Kimi K2.5 measured |
| LMArena Text: non-English | Not measured | 1433 | Only Kimi K2.5 measured |
| LMArena Text: Chinese | Not measured | 1494 | Only Kimi K2.5 measured |
| LMArena Text: German | Not measured | 1441 | Only Kimi K2.5 measured |
| LMArena Text: Japanese | Not measured | 1423 | Only Kimi K2.5 measured |
| LMArena Text: Korean | Not measured | 1412 | Only Kimi K2.5 measured |
| LMArena Text: French | Not measured | 1454 | Only Kimi K2.5 measured |
| LMArena Text: Spanish | Not measured | 1450 | Only Kimi K2.5 measured |
| LMArena Text: Russian | Not measured | 1435 | Only Kimi K2.5 measured |
| LMArena Text: Polish | Not measured | 1445 | Only Kimi K2.5 measured |
| LMArena Text: multi-turn | Not measured | 1444 | Only Kimi K2.5 measured |
| LMArena Text: longer queries | Not measured | 1445 | Only Kimi K2.5 measured |
| LMArena Text: writing and language | Not measured | 1425 | Only Kimi K2.5 measured |
| LMArena Text: entertainment and media | Not measured | 1417 | Only Kimi K2.5 measured |
| LMArena Vision: English | Not measured | 1274 | Only Kimi K2.5 measured |
| LMArena Vision: Chinese | Not measured | 1328 | Only Kimi K2.5 measured |
| LMArena Vision: reading text | Not measured | 1276 | Only Kimi K2.5 measured |
| LMArena Vision: humor | Not measured | 1284 | Only Kimi K2.5 measured |
| LMArena Vision: captioning | Not measured | 1257 | Only Kimi K2.5 measured |
| LMArena Vision: recognizing things | Not measured | 1285 | Only Kimi K2.5 measured |
| LMArena Vision: creative writing | Not measured | 1269 | Only Kimi K2.5 measured |
| Neither model has a result on the other 12 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | DeepSeek V4 Flash 0731 | 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
DeepSeek V4 Flash 0731 costs less
Repository review
50K input + 3K output tokens
DeepSeek V4 Flash 0731 costs less
Agent loop
200K input + 10K output tokens
DeepSeek V4 Flash 0731 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 | DeepSeek V4 Flash 0731 | MoonshotAI Kimi K2.5 |
|---|---|---|
| Maker | DeepSeek | Moonshot AI |
| Weights | Open weights | Open weights |
| Released | July 31, 2026 | January 27, 2026 |
| Context window | 1.05M | 262K |
| Input $/Mtok | $0.14 | $0.532 |
| Output $/Mtok | $0.28 | $2.85 |
| Coding index | 51.8 | 38.6 |
| Independent measurements behind it | 4 | 3 |
| Cost per solved task | $0.01 | $0.24 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V4 Flash 0731 better than Kimi K2.5 for coding?
Pick DeepSeek V4 Flash 0731 for coding: it scores 51.8 on the Vibe Coding Index against 38.6 for Kimi K2.5. Pick Kimi K2.5 for work measured by CorpFin (Vals AI) or TaxEval (Vals AI), where it leads.
Which is cheaper, DeepSeek V4 Flash 0731 or Kimi K2.5?
DeepSeek V4 Flash 0731 costs $0.14 per million input tokens and $0.28 per million output tokens; Kimi K2.5 costs $0.532 and $2.85. Per solved SWE-bench Verified task, DeepSeek V4 Flash 0731 costs $0.01 and Kimi K2.5 $0.24.
Which has the bigger context window, DeepSeek V4 Flash 0731 or Kimi K2.5?
DeepSeek V4 Flash 0731, with 1.05M tokens against 262K.
How many benchmarks is the DeepSeek V4 Flash 0731 vs Kimi K2.5 comparison based on?
34 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4 Flash 0731's coding index rests on 4 independent benchmarks and Kimi K2.5's on 3. Where only one model has a result, the page says so instead of naming a winner.