# DeepSeek-V3.2-Exp vs o4-mini

> DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 41.6 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v3-2-exp-vs-o4-mini
- Last updated: 2026-10-11
- Shared benchmarks: 35

## Summary

- They share 35 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 7 categories and o4-mini in 2 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-V3.2-Exp leads 62.4 to 54.0.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 70% for DeepSeek-V3.2-Exp and 45% for o4-mini.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- o4-mini accepts more context: 200K tokens versus 164K.
- DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 44.3 | 41.6 |
| Rank | 78 | 132 |
| Context | 164K | 200K |
| Input $/M | $0.26 | $1.10 |
| Output $/M | $0.38 | $4.40 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek-V3.2-Exp: 46.5 (#65)
- o4-mini: 40.9 (#127)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| SWE-bench Verified (bash only) | 70% | 45% |
| Aider Polyglot | 74.2% | 72% |
| WeirdML | 39.5% | 52.6% |
| LMArena Coding | 1454 | 1368 |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| GSO | — | 3.6% |
| CadEval | — | 62% |
| ALE-Bench | — | 826.17 |
| AlgoTune | — | 1.72 |

## Agentic & Tool Use

- DeepSeek-V3.2-Exp: 32.7 (#59)
- o4-mini: 32.6 (#61)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | 56.7% | 53.2% |
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| GDPval | — | 25.3% |
| TheAgentCompany | 42.9% | — |
| METR Time Horizons | — | 63.9% |
| Vending-Bench 2 | 1,034 | — |

## Reasoning

- DeepSeek-V3.2-Exp: 22.1 (#208)
- o4-mini: 24.6 (#162)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| ARC-AGI-2 | 4% | 6.1% |
| Kagi LLM Benchmark | 52.2% | 67.6% |
| ARC-AGI-1 | 57% | 58.7% |
| CritPt | 2.9% | 0.6% |
| Chess Puzzles | 14% | 26% |
| LMArena Hard Prompts | 1434 | 1351 |
| DTBench | 87.7% | 77.6% |
| LMCA | 29.1% | 26.5% |
| Epoch Capabilities Index | 146.27 | 145.64 |
| SimpleBench | — | 38.7% |
| NYT Connections (extended) | 36.7% | — |
| EnigmaEval | — | 9.2% |
| Thematic Generalization | 65% | — |
| Mystery Game Puzzles | — | 5% |
| ForecastBench | — | 61.8 |

## Math

- DeepSeek-V3.2-Exp: 41.7 (#87)
- o4-mini: 40.8 (#89)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 81.7% |
| LMArena Math | 1435 | 1389 |
| FrontierMath (Feb 2025 set) | 22.1% | 24.8% |
| FrontierMath Tier 4 (v1) | 2.1% | 6.3% |
| FrontierMath (Tiers 1-3) | — | 36.1% |
| FrontierMath Tier 4 | — | 4.9% |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| Omni-MATH | — | 72% |
| MATH Level 5 | — | 97.8% |

## Knowledge

- DeepSeek-V3.2-Exp: 51.7 (#66)
- o4-mini: 43.6 (#91)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| GPQA Diamond | 83.4% | 79.6% |
| Vectara Hallucination Rate | 5.3% | 18.6% |
| LMArena Expert | 1436 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| SimpleQA Verified | — | 19.6% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| GPQA (HELM) | — | 73.5% |

## Multimodal

- DeepSeek-V3.2-Exp: —
- o4-mini: 40.2 (#49)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| LMArena Vision | — | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |

## Multilingual

- DeepSeek-V3.2-Exp: 52.2 (#90)
- o4-mini: 47.0 (#154)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| LMArena Non-English | 1409 | 1337 |
| LMArena Chinese | 1461 | 1354 |
| LMArena French | 1433 | 1364 |
| LMArena German | 1440 | 1336 |
| LMArena Japanese | 1374 | 1308 |
| LMArena Korean | 1371 | 1312 |
| LMArena Russian | 1424 | 1334 |
| LMArena Spanish | 1440 | 1347 |

## Instruction Following

- DeepSeek-V3.2-Exp: 74.5 (#93)
- o4-mini: 75.2 (#68)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1413 | 1321 |
| IFEval | — | 92.8% |

## Long Context

- DeepSeek-V3.2-Exp: 47.6 (#16)
- o4-mini: 45.5 (#33)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| Fiction.LiveBench | 83.3% | 77.8% |
| LMArena Longer Query | 1428 | 1315 |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |

## Writing & Preference

- DeepSeek-V3.2-Exp: 62.4 (#77)
- o4-mini: 54.0 (#152)

| Benchmark | DeepSeek-V3.2-Exp | o4-mini |
|---|---|---|
| LMArena Text | 1425 | 1353 |
| LMArena Creative Writing | 1403 | 1294 |
| LMArena Multi-Turn | 1427 | 1350 |
| Short-Story Creative Writing | — | 75% |
| EQ-Bench Creative Writing | 1515 | — |
| WildBench | — | 85.4% |

## FAQ

### Is DeepSeek-V3.2-Exp better than o4-mini?

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 41.6 on the Noometry Index.

### Which is cheaper, DeepSeek-V3.2-Exp or o4-mini?

DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; o4-mini lists at $1.10 and $4.40.

### Is DeepSeek-V3.2-Exp or o4-mini better for coding?

DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 40.9 in the Noometry coding category.

### Which has the bigger context window?

o4-mini does, with 200K tokens against 164K.

### How many benchmarks do DeepSeek-V3.2-Exp and o4-mini share?

35 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and o4-mini has 60.