Model comparison
DeepSeek V4 Flash vs GPT-5 Nano
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 33.5 on the Noometry Index. GPT-5 Nano costs 1.9× less per token, which makes it the better buy when DeepSeek V4 Flash's lead doesn't matter for your workload.
Last verified . 33 shared benchmarks.
DeepSeekSummary
- They share 33 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 7 categories and GPT-5 Nano in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 16.3.
- The biggest single-benchmark swing is ARC-AGI-1: 89% for DeepSeek V4 Flash and 20.7% for GPT-5 Nano.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $0.15 / $0.60 for DeepSeek V4 Flash.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 400K.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Flash | GPT-5 Nano | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 53.6 | 33.5 |
| Released | 2026-04-24 | 2025-08-07 |
| Weights | Open | Proprietary |
| Context window | 1M | 400K |
| Max output | 393K | 128K |
| Input $ / M tokens | $0.15 | $0.05 |
| Output $ / M tokens | $0.60 | $0.40 |
| Results tracked | 41 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Flash leads
DeepSeek V4 Flash: 47.9 (#59), GPT-5 Nano: 33.6 (#254)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| WeirdML | 63% | 38.1% |
| LMArena Coding | 1457 | 1351 |
| ALE-Bench | 1,306 | 718.67 |
| FrontierCode | 18.8% | — |
| SWE-bench Verified (bash only) | — | 34.8% |
| LMArena WebDev | 1582 | — |
| SciCode | 49.9% | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Flash: —, GPT-5 Nano: 25.8 (#106)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| Terminal-Bench | — | 21.8% |
| Berkeley Function Calling Leaderboard | — | 51.5% |
Reasoning DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.7 (#30), GPT-5 Nano: 16.3 (#306)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| ARC-AGI-2 | 61.4% | 2.6% |
| Kagi LLM Benchmark | 52.2% | 62.2% |
| ARC-AGI-1 | 89% | 20.7% |
| Chess Puzzles | 33% | 27% |
| LMArena Hard Prompts | 1444 | 1328 |
| Mystery Game Puzzles | 34% | 9% |
| DTBench | 90.9% | 62.7% |
| LMCA | 41.7% | 7.9% |
| Epoch Capabilities Index | 154.49 | 139.38 |
| SimpleBench | 61.1% | — |
| NYT Connections (extended) | 89.6% | — |
| CritPt | 16.6% | — |
| ForecastBench | — | 59.1 |
Math DeepSeek V4 Flash leads
DeepSeek V4 Flash: 60.3 (#37), GPT-5 Nano: 29.4 (#241)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.5% | 20% |
| FrontierMath Tier 4 | 24.4% | 2.4% |
| OTIS Mock AIME 2024-2025 | 94.4% | 81.1% |
| ProofBench | 56% | 12% |
| LMArena Math | 1427 | 1317 |
| MathArena Final-Answer Competitions | 76.5% | — |
| Omni-MATH | — | 54.6% |
| MATH Level 5 | — | 95.2% |
| FrontierMath (Feb 2025 set) | — | 8.3% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge DeepSeek V4 Flash leads
DeepSeek V4 Flash: 55.4 (#48), GPT-5 Nano: 35.9 (#178)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| GPQA Diamond | 91% | 69.4% |
| SimpleQA Verified | 33.6% | 11.7% |
| LMArena Expert | 1441 | 1321 |
| MMLU-Pro | — | 77.8% |
| Vectara Hallucination Rate | — | 10.5% |
| GPQA (HELM) | — | 67.9% |
Multimodal Not comparable
DeepSeek V4 Flash: —, GPT-5 Nano: 31.3 (#108)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| LMArena Vision | — | 1159 |
| VPCT | — | 37.2% |
Multilingual DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.0 (#72), GPT-5 Nano: 45.3 (#172)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| LMArena Non-English | 1420 | 1313 |
| LMArena Chinese | 1468 | 1356 |
| LMArena German | 1418 | 1327 |
| LMArena Japanese | 1406 | 1226 |
| LMArena Korean | 1384 | 1269 |
| LMArena Russian | 1428 | 1296 |
| LMArena Spanish | 1436 | 1360 |
| LMArena French | 1439 | — |
Instruction Following Too close to call
DeepSeek V4 Flash: 74.9 (#81), GPT-5 Nano: 75.0 (#79)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| LMArena Instruction Following | 1421 | 1306 |
| IFEval | — | 93.2% |
Long Context DeepSeek V4 Flash leads
DeepSeek V4 Flash: 43.8 (#85), GPT-5 Nano: 31.3 (#281)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| LMArena Longer Query | 1434 | 1312 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference DeepSeek V4 Flash leads
DeepSeek V4 Flash: 63.8 (#61), GPT-5 Nano: 39.1 (#249)
| Benchmark | DeepSeek V4 Flash | GPT-5 Nano |
|---|---|---|
| LMArena Text | 1432 | 1320 |
| LMArena Creative Writing | 1403 | 1249 |
| EQ-Bench Creative Writing | 1559 | 705 |
| LMArena Multi-Turn | 1449 | 1311 |
| WildBench | — | 80.6% |
Frequently asked questions
Is DeepSeek V4 Flash better than GPT-5 Nano?
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 33.5 on the Noometry Index. GPT-5 Nano costs 1.9× less per token, which makes it the better buy when DeepSeek V4 Flash's lead doesn't matter for your workload.
Which is cheaper, DeepSeek V4 Flash or GPT-5 Nano?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; DeepSeek V4 Flash lists at $0.15 and $0.60.
Is DeepSeek V4 Flash or GPT-5 Nano better for coding?
DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 33.6 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Flash does, with 1M tokens against 400K.
How many benchmarks do DeepSeek V4 Flash and GPT-5 Nano share?
33 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and GPT-5 Nano has 49.