Model comparison
DeepSeek-V3.2-Speciale vs gpt-oss-120b
DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.1 to 36.0 on the Noometry Index. gpt-oss-120b costs 12× less per token, which makes it the better buy when DeepSeek-V3.2-Speciale's lead doesn't matter for your workload.
Last verified . 3 shared benchmarks.
DeepSeekSummary
- They share 3 benchmarks with published results for both. DeepSeek-V3.2-Speciale scores higher in 2 categories and gpt-oss-120b in 1 category; 2 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V3.2-Speciale leads 32.1 to 19.5.
- The biggest single-benchmark swing is SimpleBench: 52.6% for DeepSeek-V3.2-Speciale and 22.1% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $0.58 / $1.68 for DeepSeek-V3.2-Speciale.
- gpt-oss-120b accepts more context: 131K tokens versus 128K.
Side by side
| DeepSeek-V3.2-Speciale | gpt-oss-120b | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 39.1 | 36.0 |
| Released | 2025-12-01 | 2025-08-05 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 128K | 41K |
| Input $ / M tokens | $0.58 | $0.037 |
| Output $ / M tokens | $1.68 | $0.17 |
| Results tracked | 3 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 39.9 (#136), gpt-oss-120b: 33.2 (#251)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| WeirdML | 46.7% | 48.2% |
| SWE-bench Verified (bash only) | — | 26% |
| Aider Polyglot | — | 41.8% |
| SciCode | — | 36% |
| LMArena Coding | — | 1380 |
| ALE-Bench | — | 575.62 |
| AlgoTune | — | 1.41 |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Speciale: —, gpt-oss-120b: 12.1 (#145)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| Terminal-Bench | — | 18.7% |
| APEX-Agents | — | 4.4% |
| METR Time Horizons | — | 56.6% |
| Vending-Bench 2 | — | -21.53 |
Reasoning DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 32.1 (#65), gpt-oss-120b: 19.5 (#236)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| SimpleBench | 52.6% | 22.1% |
| Kagi LLM Benchmark | — | 58.6% |
| CritPt | — | 1.1% |
| Chess Puzzles | — | 20% |
| LMArena Hard Prompts | — | 1364 |
| Mystery Game Puzzles | — | 2% |
| DTBench | — | 76.3% |
| LMCA | — | 22.1% |
| Surface Evolver Bench | — | 25% |
| Epoch Capabilities Index | — | 139.94 |
Math Not comparable
DeepSeek-V3.2-Speciale: —, gpt-oss-120b: 50.6 (#43)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| Omni-MATH | — | 68.8% |
| LMArena Math | — | 1388 |
Knowledge Not comparable
DeepSeek-V3.2-Speciale: —, gpt-oss-120b: 42.3 (#89)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | — | 75.8% |
| MMLU-Pro | — | 79.5% |
| Confabulations | — | 15.7% |
| Vectara Hallucination Rate | — | 14.2% |
| GPQA (HELM) | — | 68.4% |
| LMArena Expert | — | 1356 |
Multilingual Not comparable
DeepSeek-V3.2-Speciale: —, gpt-oss-120b: 48.2 (#142)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | — | 1350 |
| LMArena Chinese | — | 1384 |
| LMArena French | — | 1370 |
| LMArena German | — | 1355 |
| LMArena Japanese | — | 1332 |
| LMArena Korean | — | 1283 |
| LMArena Russian | — | 1343 |
| LMArena Spanish | — | 1388 |
Instruction Following Not comparable
DeepSeek-V3.2-Speciale: —, gpt-oss-120b: 69.5 (#168)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| IFEval | — | 83.6% |
| LMArena Instruction Following | — | 1318 |
Long Context Not comparable
DeepSeek-V3.2-Speciale: —, gpt-oss-120b: 31.4 (#273)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| Fiction.LiveBench | — | 44.4% |
| LMArena Longer Query | — | 1319 |
Writing & Preference Too close to call
DeepSeek-V3.2-Speciale: 46.1 (#217), gpt-oss-120b: 46.7 (#213)
| Benchmark | DeepSeek-V3.2-Speciale | gpt-oss-120b |
|---|---|---|
| EQ-Bench Creative Writing | 1276 | 961 |
| LMArena Text | — | 1366 |
| LMArena Creative Writing | — | 1274 |
| Short-Story Creative Writing | — | 77.1% |
| WildBench | — | 84.5% |
| LMArena Multi-Turn | — | 1340 |
Frequently asked questions
Is DeepSeek-V3.2-Speciale better than gpt-oss-120b?
DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.1 to 36.0 on the Noometry Index. gpt-oss-120b costs 12× less per token, which makes it the better buy when DeepSeek-V3.2-Speciale's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.2-Speciale or gpt-oss-120b?
gpt-oss-120b is cheaper. It lists at $0.037 per million input tokens and $0.17 per million output tokens; DeepSeek-V3.2-Speciale lists at $0.58 and $1.68.
Is DeepSeek-V3.2-Speciale or gpt-oss-120b better for coding?
DeepSeek-V3.2-Speciale scores higher on coding benchmarks: 39.9 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
gpt-oss-120b does, with 131K tokens against 128K.
How many benchmarks do DeepSeek-V3.2-Speciale and gpt-oss-120b share?
3 benchmarks have published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and gpt-oss-120b has 48.