DeepSeek V3.2 vs GPT-5 Nano
DeepSeek V3.2 leads GPT-5 Nano on the Artificial Analysis Intelligence Index, 21.5 to 13.0, and GPT-5 Nano costs 2.2x less per output token ($0.40 vs $0.88 per million). Updated October 22, 2026; 60 benchmarks have a result for both.
DeepSeek V3.2 or GPT-5 Nano: which to pick
Only DeepSeek V3.2 has a coding score, 30.0 on the Vibe Coding Index; GPT-5 Nano has none, so the two cannot be ranked on coding.
Pick GPT-5 Nano for output tokens 2.2x cheaper ($0.40 vs $0.88 per million) and a lower cost per solved task ($0.109 against $0.99).
- Shared results
- 60
- DeepSeek V3.2 only
- 15
- GPT-5 Nano only
- 27
- 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
No call
Only DeepSeek V3.2 has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
DeepSeek V3.2
DeepSeek V3.2 leads, 21.5 against 13.0.
Same board, both measured
Reasoning
GPQA Diamond
DeepSeek V3.2
DeepSeek V3.2 leads, 80.3% against 63.4%.
Same board, both measured
Math
AIME
DeepSeek V3.2
DeepSeek V3.2 leads, 84.6% against 81.2%.
Same board, both measured
Knowledge
MMLU Pro
DeepSeek V3.2
DeepSeek V3.2 leads, 84.9% against 76.1%.
Same board, both measured
Web design
Design Arena: websites
DeepSeek V3.2
DeepSeek V3.2 leads, 1175 against 1101.
Same board, both measured
Long documents
Listed context window
GPT-5 Nano
GPT-5 Nano takes 400K tokens against 164K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
GPT-5 Nano
DeepSeek V3.2 $0.0007, GPT-5 Nano $0.0003 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-5 Nano
DeepSeek V3.2 $0.99, GPT-5 Nano $0.109.
Published attempt cost and pass rate
At a glance
- Coding index
- DeepSeek V3.2 30.0 from 3 benchmarks, GPT-5 Nano not benchmarked.
- Widest category
- Agentic work, 54.1 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: DeepSeek V3.2 90.6% against GPT-5 Nano 36.5%, 54.1 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V3.2 $0.88 against GPT-5 Nano $0.40 per million output tokens, 2.2x apart.
- Context window
- DeepSeek V3.2 takes 164K tokens, GPT-5 Nano 400K.
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 V3.2 | Bars | GPT-5 Nano | Ties |
|---|---|---|---|---|
| Coding | 7 | 0 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 1 | 0 | ||
| Reasoning | 8 | 0 | ||
| Math | 5 | 1 | ||
| Knowledge | 10 | 0 | ||
| Web design | 6 | 0 | ||
| General | 19 | 1 | ||
| All shared boards | 58 | 2 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index70
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| SWE-bench Verified | 70.0% | 34.8% | DeepSeek V3.2 |
| SWE-bench Multilingual | 59.0% | Not measured | Only DeepSeek V3.2 measured |
| LMArena WebDev | 1362 | Not measured | Only DeepSeek V3.2 measured |
| AA Coding Index | 44.2 | Not measured | Only DeepSeek V3.2 measured |
| LiveCodeBench (Vals AI) | 80.7% | 70.2% | DeepSeek V3.2 |
| SWE-bench Verified (Vals AI) | 67.6% | Not measured | Only DeepSeek V3.2 measured |
| Vibe Code Bench (Vals AI) | 5.1% | Not measured | Only DeepSeek V3.2 measured |
| LiveCodeBench (Artificial Analysis run) | 86.2% | 78.9% | DeepSeek V3.2 |
| Terminal-Bench Hard (Artificial Analysis run) | 35.6% | 17.4% | DeepSeek V3.2 |
| Terminal-Bench | 39.6% | 21.8% | DeepSeek V3.2 |
| WeirdML | Not measured | 38.1% | Only GPT-5 Nano measured |
| LMArena Text: coding | 1453 | 1351 | DeepSeek V3.2 |
| LMArena Text: software and IT | 1449 | 1345 | DeepSeek V3.2 |
| LMArena WebDev: React | 1350 | Not measured | Only DeepSeek V3.2 measured |
| LMArena WebDev: HTML | 1371 | Not measured | Only DeepSeek V3.2 measured |
| Neither model has a result on the other 24 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| AA Intelligence Index | 21.5 | 13.0 | DeepSeek V3.2 |
| Epoch Capabilities Index | 146.1 | 139.2 | DeepSeek V3.2 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 90.6% | 36.5% | DeepSeek V3.2 |
| APEX-Agents | 21.3% | Not measured | Only DeepSeek V3.2 measured |
| Vending-Bench 2 | $1,034 | Not measured | Only DeepSeek V3.2 measured |
| Berkeley Function Calling Leaderboard (BFCL) | Not measured | 51.4% | Only GPT-5 Nano measured |
| Neither model has a result on the other 37 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index80
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| GPQA Diamond | 80.3% | 63.4% | DeepSeek V3.2 |
| MMMU Pro (Vals AI) | Not measured | 70.9% | Only GPT-5 Nano measured |
| GPQA Diamond (Artificial Analysis run) | 84.0% | 67.6% | DeepSeek V3.2 |
| Humanity's Last Exam (Artificial Analysis run) | 24.6% | 9.5% | DeepSeek V3.2 |
| GPQA Diamond (Epoch AI run) | Not measured | 69.4% | Only GPT-5 Nano measured |
| ARC-AGI | 57.0% | 20.7% | DeepSeek V3.2 |
| ARC-AGI-2 | 4.0% | 2.6% | DeepSeek V3.2 |
| Chess puzzles | Not measured | 22.0% | Only GPT-5 Nano measured |
| Mystery game puzzles | Not measured | 9.0% | Only GPT-5 Nano measured |
| VPCT | Not measured | 37.2% | Only GPT-5 Nano measured |
| LMArena Text: hard prompts | 1434 | 1328 | DeepSeek V3.2 |
| LMArena Text: expert | 1436 | 1321 | DeepSeek V3.2 |
| LMArena Text: hard prompts in English | 1446 | 1338 | DeepSeek V3.2 |
| LMArena Vision: diagrams | Not measured | 1141 | Only GPT-5 Nano measured |
| LMArena Vision: homework | Not measured | 1157 | Only GPT-5 Nano measured |
| Kagi LLM Benchmark | Not measured | 62.2% | Only GPT-5 Nano measured |
| Neither model has a result on the other 6 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index51
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| AIME | 84.6% | 81.2% | DeepSeek V3.2 |
| MATH 500 (Vals AI) | Not measured | 93.8% | Only GPT-5 Nano measured |
| MGSM (Vals AI) | 86.0% | 89.3% | GPT-5 Nano |
| AA Math Index | 92.0 | 83.7 | DeepSeek V3.2 |
| AIME 2025 (Artificial Analysis run) | 92.0% | 83.7% | DeepSeek V3.2 |
| OTIS Mock AIME | Not measured | 81.1% | Only GPT-5 Nano measured |
| FrontierMath, tiers 1 to 3 | Not measured | 20.0% | Only GPT-5 Nano measured |
| FrontierMath, tier 4 | Not measured | 2.4% | Only GPT-5 Nano measured |
| ProofBench | Not measured | 12.0% | Only GPT-5 Nano measured |
| MATH level 5 | Not measured | 95.2% | Only GPT-5 Nano measured |
| LMArena Text: math | 1435 | 1317 | DeepSeek V3.2 |
| LMArena Text: mathematical work | 1440 | 1326 | DeepSeek V3.2 |
| Neither model has a result on the other 1 board here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index100
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| MMLU Pro | 84.9% | 76.1% | DeepSeek V3.2 |
| LegalBench (Vals AI) | 76.1% | 50.1% | DeepSeek V3.2 |
| CaseLaw (Vals AI) | 55.4% | 52.6% | DeepSeek V3.2 |
| CorpFin (Vals AI) | 51.0% | Not measured | Only DeepSeek V3.2 measured |
| TaxEval (Vals AI) | 68.2% | 67.4% | DeepSeek V3.2 |
| MortgageTax (Vals AI) | Not measured | 53.6% | Only GPT-5 Nano measured |
| MedQA (Vals AI) | 93.9% | 93.3% | DeepSeek V3.2 |
| MedCode (Vals AI) | Not measured | 30.4% | Only GPT-5 Nano measured |
| MMLU Pro (Artificial Analysis run) | 86.2% | 78.0% | DeepSeek V3.2 |
| SimpleQA Verified | Not measured | 11.7% | Only GPT-5 Nano measured |
| LMArena Text: science | 1446 | 1318 | DeepSeek V3.2 |
| LMArena Text: medicine and healthcare | 1440 | 1335 | DeepSeek V3.2 |
| LMArena Text: legal and government | 1425 | 1298 | DeepSeek V3.2 |
| LMArena Text: business and finance | 1410 | 1311 | DeepSeek V3.2 |
| Neither model has a result on the other 4 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index60
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| Design Arena: websites | 1175 | 1101 | DeepSeek V3.2 |
| Design Arena: UI components | 1146 | 1064 | DeepSeek V3.2 |
| Design Arena: game dev | 1131 | 1044 | DeepSeek V3.2 |
| Design Arena: data viz | 1158 | 1056 | DeepSeek V3.2 |
| Design Arena: 3D | 1132 | 965 | DeepSeek V3.2 |
| Design Arena: all code categories | 1163 | 1084 | DeepSeek V3.2 |
| Design Arena: SVG | 1032 | Not measured | Only DeepSeek V3.2 measured |
| Design Arena: ASCII art | 1083 | Not measured | Only DeepSeek V3.2 measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index191
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | Leader |
|---|---|---|---|
| SAGE (Vals AI) | Not measured | 30.4% | Only GPT-5 Nano measured |
| MedScribe (Vals AI) | Not measured | 72.9% | Only GPT-5 Nano measured |
| IFBench | 60.7% | 67.6% | GPT-5 Nano |
| Long context reasoning (AA-LCR) | 73.3% | 45.0% | DeepSeek V3.2 |
| DTBench | 85.6% | 62.7% | DeepSeek V3.2 |
| LMCA | 28.8% | 7.9% | DeepSeek V3.2 |
| Fiction.LiveBench | Not measured | 44.4% | Only GPT-5 Nano measured |
| CL-bench | 12.4% | Not measured | Only DeepSeek V3.2 measured |
| CL-bench Life | 7.4% | Not measured | Only DeepSeek V3.2 measured |
| LMArena Text | 1425 | 1320 | DeepSeek V3.2 |
| LMArena Text: instruction following | 1413 | 1306 | DeepSeek V3.2 |
| LMArena Text: creative writing | 1400 | 1249 | DeepSeek V3.2 |
| LMArena Vision | Not measured | 1159 | Only GPT-5 Nano measured |
| LMArena Text: English | 1439 | 1321 | DeepSeek V3.2 |
| LMArena Text: non-English | 1409 | 1313 | DeepSeek V3.2 |
| LMArena Text: Chinese | 1455 | 1356 | DeepSeek V3.2 |
| LMArena Text: German | 1413 | 1327 | DeepSeek V3.2 |
| LMArena Text: Japanese | 1374 | 1227 | DeepSeek V3.2 |
| LMArena Text: Korean | 1370 | 1269 | DeepSeek V3.2 |
| LMArena Text: French | 1429 | Not measured | Only DeepSeek V3.2 measured |
| LMArena Text: Spanish | 1424 | 1360 | DeepSeek V3.2 |
| LMArena Text: Russian | 1424 | 1296 | DeepSeek V3.2 |
| LMArena Text: Polish | 1419 | 1314 | DeepSeek V3.2 |
| LMArena Text: multi-turn | 1427 | 1311 | DeepSeek V3.2 |
| LMArena Text: longer queries | 1428 | 1312 | DeepSeek V3.2 |
| LMArena Text: writing and language | 1407 | 1283 | DeepSeek V3.2 |
| LMArena Text: entertainment and media | 1395 | 1268 | DeepSeek V3.2 |
| LMArena Vision: English | Not measured | 1170 | Only GPT-5 Nano measured |
| LMArena Vision: Chinese | Not measured | 1176 | Only GPT-5 Nano measured |
| LMArena Vision: reading text | Not measured | 1152 | Only GPT-5 Nano measured |
| LMArena Vision: humor | Not measured | 1110 | Only GPT-5 Nano measured |
| Neither model has a result on the other 15 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | DeepSeek V3.2 | GPT-5 Nano | 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
GPT-5 Nano costs less
Repository review
50K input + 3K output tokens
GPT-5 Nano costs less
Agent loop
200K input + 10K output tokens
GPT-5 Nano 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 V3.2 | OpenAI GPT-5 Nano |
|---|---|---|
| Maker | DeepSeek | OpenAI |
| Weights | Open weights | No public weights listed |
| Released | December 1, 2025 | August 7, 2025 |
| Context window | 164K | 400K |
| Input $/Mtok | $0.29 | $0.05 |
| Output $/Mtok | $0.88 | $0.40 |
| Coding index | 30.0 | Not benchmarked |
| Independent measurements behind it | 3 | 0 |
| Cost per solved task | $0.99 | $0.109 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V3.2 better than GPT-5 Nano for coding?
Only DeepSeek V3.2 has a coding score, 30.0 on the Vibe Coding Index; GPT-5 Nano has none, so the two cannot be ranked on coding. Pick GPT-5 Nano for output tokens 2.2x cheaper ($0.40 vs $0.88 per million) and a lower cost per solved task ($0.109 against $0.99).
Which is cheaper, DeepSeek V3.2 or GPT-5 Nano?
DeepSeek V3.2 costs $0.29 per million input tokens and $0.88 per million output tokens; GPT-5 Nano costs $0.05 and $0.40. Per solved SWE-bench Verified task, DeepSeek V3.2 costs $0.99 and GPT-5 Nano $0.109.
Which has the bigger context window, DeepSeek V3.2 or GPT-5 Nano?
GPT-5 Nano, with 400K tokens against 164K.
How many benchmarks is the DeepSeek V3.2 vs GPT-5 Nano comparison based on?
60 benchmarks have a result for both models as of October 22, 2026. DeepSeek V3.2's coding index rests on 3 independent benchmarks and GPT-5 Nano's on 0. Where only one model has a result, the page says so instead of naming a winner.