Model comparison
DeepSeek-V3.2-Exp vs gpt-oss-20b
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 32.5 on the Noometry Index. gpt-oss-20b costs 8.1× less per token, which makes it the better buy when DeepSeek-V3.2-Exp's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
DeepSeekSummary
- They share 28 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 9 categories and gpt-oss-20b in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-V3.2-Exp leads 62.4 to 35.5.
- The biggest single-benchmark swing is Terminal-Bench: 39.6% for DeepSeek-V3.2-Exp and 3.4% for gpt-oss-20b.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.26 / $0.38 for DeepSeek-V3.2-Exp.
- DeepSeek-V3.2-Exp accepts more context: 164K tokens versus 131K.
Side by side
| DeepSeek-V3.2-Exp | gpt-oss-20b | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 44.3 | 32.5 |
| Released | 2025-09-29 | 2025-08-05 |
| Weights | Open | Open |
| Context window | 164K | 131K |
| Max output | 66K | 16K |
| Input $ / M tokens | $0.26 | $0.018 |
| Output $ / M tokens | $0.38 | $0.09 |
| Results tracked | 49 | 34 |
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-oss-20b: 37.6 (#192)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| SciCode | 38.9% | 34.4% |
| WeirdML | 39.5% | 40.9% |
| LMArena Coding | 1454 | 1306 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| ALE-Bench | — | 566.05 |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 32.7 (#59), gpt-oss-20b: 9.3 (#154)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| Terminal-Bench | 39.6% | 3.4% |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 22.1 (#208), gpt-oss-20b: 19.3 (#261)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 53.2% |
| CritPt | 2.9% | 1.4% |
| Chess Puzzles | 14% | 4% |
| LMArena Hard Prompts | 1434 | 1274 |
| DTBench | 87.7% | 68% |
| LMCA | 29.1% | 14.5% |
| Epoch Capabilities Index | 146.27 | 137.82 |
| ARC-AGI-2 | 4% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| Thematic Generalization | 65% | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), gpt-oss-20b: 39.4 (#103)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 65.3% |
| LMArena Math | 1435 | 1317 |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| Omni-MATH | — | 56.5% |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), gpt-oss-20b: 34.6 (#195)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| GPQA Diamond | 83.4% | 60.8% |
| LMArena Expert | 1436 | 1258 |
| MMLU-Pro | — | 74% |
| Vectara Hallucination Rate | 5.3% | — |
| GPQA (HELM) | — | 59.4% |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.2 (#90), gpt-oss-20b: 42.2 (#197)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| LMArena Non-English | 1409 | 1268 |
| LMArena Chinese | 1461 | 1314 |
| LMArena German | 1440 | 1255 |
| LMArena Japanese | 1374 | 1244 |
| LMArena Korean | 1371 | 1236 |
| LMArena Russian | 1424 | 1278 |
| LMArena Spanish | 1440 | 1267 |
| LMArena French | 1433 | — |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), gpt-oss-20b: 61.8 (#240)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| LMArena Instruction Following | 1413 | 1236 |
| IFEval | — | 73.2% |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), gpt-oss-20b: 37.9 (#209)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| LMArena Longer Query | 1428 | 1250 |
| 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-oss-20b: 35.5 (#265)
| Benchmark | DeepSeek-V3.2-Exp | gpt-oss-20b |
|---|---|---|
| LMArena Text | 1425 | 1287 |
| LMArena Creative Writing | 1403 | 1201 |
| EQ-Bench Creative Writing | 1515 | 666 |
| LMArena Multi-Turn | 1427 | 1268 |
| WildBench | — | 73.7% |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than gpt-oss-20b?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 32.5 on the Noometry Index. gpt-oss-20b costs 8.1× less per token, which makes it the better buy when DeepSeek-V3.2-Exp's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.2-Exp or gpt-oss-20b?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; DeepSeek-V3.2-Exp lists at $0.26 and $0.38.
Is DeepSeek-V3.2-Exp or gpt-oss-20b better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 37.6 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-V3.2-Exp does, with 164K tokens against 131K.
How many benchmarks do DeepSeek-V3.2-Exp and gpt-oss-20b share?
28 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and gpt-oss-20b has 34.