Model comparison
DeepSeek-V3.2-Exp vs GPT-5.1
GPT-5.1 is the stronger model overall, scoring 49.0 to 44.3 on the Noometry Index. DeepSeek-V3.2-Exp costs 12× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.
Last verified . 37 shared benchmarks.
DeepSeekSummary
- They share 37 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 3 categories and GPT-5.1 in 6 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.1 leads 39.8 to 22.1.
- The biggest single-benchmark swing is WeirdML: 39.5% for DeepSeek-V3.2-Exp and 60.8% for GPT-5.1.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $1.25 / $10 for GPT-5.1.
- GPT-5.1 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-5.1 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 44.3 | 49.0 |
| Released | 2025-09-29 | 2025-11-13 |
| Weights | Open | Proprietary |
| Context window | 164K | 400K |
| Max output | 66K | 128K |
| Input $ / M tokens | $0.26 | $1.25 |
| Output $ / M tokens | $0.38 | $10 |
| Results tracked | 49 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-V3.2-Exp: 46.5 (#65), GPT-5.1: 46.4 (#66)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| SWE-bench Verified (bash only) | 70% | 66% |
| LMArena WebDev | 1362 | 1395 |
| SciCode | 38.9% | 43.3% |
| WeirdML | 39.5% | 60.8% |
| LMArena Coding | 1454 | 1454 |
| SWE-bench Verified | — | 68% |
| Aider Polyglot | 74.2% | — |
| SWE-bench Multilingual | 59% | — |
| GSO | — | 13.7% |
| LiveBench Coding | — | 72.5% |
| ALE-Bench | — | 1,192 |
Agentic & Tool Use Too close to call
DeepSeek-V3.2-Exp: 32.7 (#59), GPT-5.1: 32.7 (#60)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| Terminal-Bench | 39.6% | 47.6% |
| Vending-Bench 2 | 1,034 | 1,473 |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| DeepResearch Bench | — | 42.8% |
| LMArena Search | — | 1199 |
Reasoning GPT-5.1 leads
DeepSeek-V3.2-Exp: 22.1 (#208), GPT-5.1: 39.8 (#58)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| ARC-AGI-2 | 4% | 17.6% |
| ARC-AGI-1 | 57% | 72.8% |
| CritPt | 2.9% | 4.9% |
| Chess Puzzles | 14% | 32% |
| LMArena Hard Prompts | 1434 | 1457 |
| DTBench | 87.7% | 90.1% |
| LMCA | 29.1% | 43.9% |
| Epoch Capabilities Index | 146.27 | 149.64 |
| SimpleBench | — | 53.2% |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| EnigmaEval | — | 11.2% |
| Thematic Generalization | 65% | — |
| LiveBench Reasoning | — | 95.8% |
| Mystery Game Puzzles | — | 19% |
| LiveBench Data Analysis | — | 72.1% |
| ForecastBench | — | 58.1 |
| LiveBench | — | 78.8% |
Math GPT-5.1 leads
DeepSeek-V3.2-Exp: 41.7 (#87), GPT-5.1: 52.2 (#51)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 88.6% |
| LMArena Math | 1435 | 1447 |
| FrontierMath (Feb 2025 set) | 22.1% | 31% |
| FrontierMath Tier 4 (v1) | 2.1% | 12.5% |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| Omni-MATH | — | 46.4% |
| LiveBench Math | — | 94.5% |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), GPT-5.1: 50.6 (#71)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| GPQA Diamond | 83.4% | 87.6% |
| Vectara Hallucination Rate | 5.3% | 10.9% |
| LMArena Expert | 1436 | 1470 |
| Humanity's Last Exam | — | 23.7% |
| SimpleQA Verified | — | 48% |
| MMLU-Pro | — | 57.9% |
| GPQA (HELM) | — | 44.2% |
Multimodal Not comparable
DeepSeek-V3.2-Exp: —, GPT-5.1: 44.8 (#19)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| LMArena Vision | — | 1250 |
| VPCT | — | 58.7% |
| LMArena Document | — | 1403 |
Multilingual GPT-5.1 leads
DeepSeek-V3.2-Exp: 52.2 (#90), GPT-5.1: 53.8 (#56)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| LMArena Non-English | 1409 | 1431 |
| LMArena Chinese | 1461 | 1495 |
| LMArena French | 1433 | 1450 |
| LMArena German | 1440 | 1438 |
| LMArena Japanese | 1374 | 1453 |
| LMArena Korean | 1371 | 1401 |
| LMArena Russian | 1424 | 1435 |
| LMArena Spanish | 1440 | 1433 |
Instruction Following GPT-5.1 leads
DeepSeek-V3.2-Exp: 74.5 (#93), GPT-5.1: 83.9 (#1)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| LMArena Instruction Following | 1413 | 1443 |
| LiveBench Instruction Following | — | 93.3% |
| IFEval | — | 93.5% |
Long Context Too close to call
DeepSeek-V3.2-Exp: 47.6 (#16), GPT-5.1: 47.6 (#14)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| CL-bench | 13.2% | 23.7% |
| CL-bench Life | 9.5% | 17.3% |
| LMArena Longer Query | 1428 | 1447 |
| Fiction.LiveBench | 83.3% | — |
Writing & Preference GPT-5.1 leads
DeepSeek-V3.2-Exp: 62.4 (#77), GPT-5.1: 64.5 (#55)
| Benchmark | DeepSeek-V3.2-Exp | GPT-5.1 |
|---|---|---|
| LMArena Text | 1425 | 1443 |
| LMArena Creative Writing | 1403 | 1427 |
| LMArena Multi-Turn | 1427 | 1450 |
| EQ-Bench Creative Writing | 1515 | — |
| WildBench | — | 86.3% |
| LiveBench Language | — | 80.2% |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than GPT-5.1?
GPT-5.1 is the stronger model overall, scoring 49.0 to 44.3 on the Noometry Index. DeepSeek-V3.2-Exp costs 12× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.2-Exp or GPT-5.1?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; GPT-5.1 lists at $1.25 and $10.
Is DeepSeek-V3.2-Exp or GPT-5.1 better for coding?
They score almost the same on coding (46.5 vs 46.4); test both on your own repository before choosing.
Which has the bigger context window?
GPT-5.1 does, with 400K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and GPT-5.1 share?
37 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and GPT-5.1 has 63.