Model comparison
DeepSeek-V3.2-Exp vs GPT-5.4 nano
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 41.9 on the Noometry Index.
Last verified . 33 shared benchmarks.
DeepSeekSummary
- They share 33 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 7 categories and GPT-5.4 nano in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 41.9.
- The biggest single-benchmark swing is Chess Puzzles: 14% for DeepSeek-V3.2-Exp and 30% for GPT-5.4 nano.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $0.20 / $1.25 for GPT-5.4 nano.
- GPT-5.4 nano accepts more context: 400K tokens versus 164K.
- DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.2-Exp | GPT-5.4 nano | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 44.3 | 41.9 |
| Released | 2025-09-29 | 2026-03-17 |
| Weights | Open | Proprietary |
| Context window | 164K | 400K |
| Max output | 66K | 128K |
| Input $ / M tokens | $0.26 | $0.20 |
| Output $ / M tokens | $0.38 | $1.25 |
| Results tracked | 49 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), GPT-5.4 nano: 43.6 (#84)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| SciCode | 38.9% | 46.9% |
| WeirdML | 39.5% | 49.2% |
| LMArena Coding | 1454 | 1405 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| ALE-Bench | — | 1,005 |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 32.7 (#59), GPT-5.4 nano: —
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning GPT-5.4 nano leads
DeepSeek-V3.2-Exp: 22.1 (#208), GPT-5.4 nano: 23.7 (#173)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| ARC-AGI-2 | 4% | 5.7% |
| Kagi LLM Benchmark | 52.2% | 39.7% |
| ARC-AGI-1 | 57% | 51.5% |
| CritPt | 2.9% | 9.3% |
| Chess Puzzles | 14% | 30% |
| LMArena Hard Prompts | 1434 | 1381 |
| DTBench | 87.7% | 80.3% |
| LMCA | 29.1% | 36.9% |
| Epoch Capabilities Index | 146.27 | 145.81 |
| NYT Connections (extended) | 36.7% | — |
| Thematic Generalization | 65% | — |
| Mystery Game Puzzles | — | 9% |
| ForecastBench | — | 57.3 |
Math Too close to call
DeepSeek-V3.2-Exp: 41.7 (#87), GPT-5.4 nano: 40.9 (#88)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 87.8% |
| ProofBench | 8% | 5% |
| LMArena Math | 1435 | 1406 |
| FrontierMath (Feb 2025 set) | 22.1% | 25.9% |
| FrontierMath Tier 4 (v1) | 2.1% | 6.3% |
| FrontierMath (Tiers 1-3) | — | 44.9% |
| FrontierMath Tier 4 | — | 12.2% |
| MathArena Final-Answer Competitions | 57.7% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), GPT-5.4 nano: 41.9 (#103)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| GPQA Diamond | 83.4% | 78.5% |
| Vectara Hallucination Rate | 5.3% | 3.1% |
| LMArena Expert | 1436 | 1396 |
| SimpleQA Verified | — | 11.7% |
Multimodal Not comparable
DeepSeek-V3.2-Exp: —, GPT-5.4 nano: 36.7 (#78)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| LMArena Vision | — | 1196 |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.2 (#90), GPT-5.4 nano: 48.6 (#140)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| LMArena Non-English | 1409 | 1359 |
| LMArena Chinese | 1461 | 1392 |
| LMArena French | 1433 | 1396 |
| LMArena German | 1440 | 1367 |
| LMArena Japanese | 1374 | 1343 |
| LMArena Korean | 1371 | 1320 |
| LMArena Russian | 1424 | 1363 |
| LMArena Spanish | 1440 | 1371 |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), GPT-5.4 nano: 71.9 (#144)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| LMArena Instruction Following | 1413 | 1362 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), GPT-5.4 nano: 41.6 (#137)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| LMArena Longer Query | 1428 | 1366 |
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), GPT-5.4 nano: 55.7 (#142)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.4 nano |
|---|---|---|
| LMArena Text | 1425 | 1372 |
| LMArena Creative Writing | 1403 | 1314 |
| LMArena Multi-Turn | 1427 | 1382 |
| EQ-Bench Creative Writing | 1515 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than GPT-5.4 nano?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 41.9 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or GPT-5.4 nano?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; GPT-5.4 nano lists at $0.20 and $1.25.
Is DeepSeek-V3.2-Exp or GPT-5.4 nano better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 43.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 nano does, with 400K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and GPT-5.4 nano share?
33 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and GPT-5.4 nano has 40.