Model comparison
DeepSeek-V2.5 (Sep 2024) vs o3-pro
o3-pro is the stronger model overall, scoring 42.9 to 37.6 on the Noometry Index.
Last verified . 1 shared benchmarks.
DeepSeekSummary
- They share 1 benchmark with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 2 categories and o3-pro in 3 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in long context, where o3-pro leads 72.2 to 39.5.
- The biggest single-benchmark swing is Aider Polyglot: 17.8% for DeepSeek-V2.5 (Sep 2024) and 84.9% for o3-pro.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | o3-pro | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 37.6 | 42.9 |
| Released | 2024-09-06 | 2025-06-10 |
| Weights | Open | Proprietary |
| Context window | — | 200K |
| Max output | — | 100K |
| Input $ / M tokens | — | $20 |
| Output $ / M tokens | — | $80 |
| Results tracked | 22 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3-pro leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), o3-pro: 55.5 (#24)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| Aider Polyglot | 17.8% | 84.9% |
| WeirdML | — | 58.2% |
| BigCodeBench Instruct | 48.6% | — |
| LMArena Coding | 1309 | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), o3-pro: 23.8 (#171)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| ARC-AGI-2 | — | 4.9% |
| Kagi LLM Benchmark | — | 72.1% |
| ARC-AGI-1 | — | 59.3% |
| LMArena Hard Prompts | 1289 | — |
| DTBench | — | 86.9% |
| LMCA | — | 38.5% |
| Epoch Capabilities Index | — | 147.42 |
Math Not comparable
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), o3-pro: —
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| LMArena Math | 1288 | — |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), o3-pro: 29.5 (#238)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| Confabulations | — | 14.2% |
| Vectara Hallucination Rate | — | 23.3% |
| LMArena Expert | 1266 | — |
Multilingual Not comparable
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), o3-pro: —
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| LMArena Non-English | 1273 | — |
| LMArena Chinese | 1318 | — |
| LMArena French | 1289 | — |
| LMArena German | 1258 | — |
| LMArena Japanese | 1228 | — |
| LMArena Korean | 1209 | — |
| LMArena Russian | 1289 | — |
| LMArena Spanish | 1248 | — |
Instruction Following Not comparable
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), o3-pro: —
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| LMArena Instruction Following | 1280 | — |
Long Context o3-pro leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), o3-pro: 72.2 (#1)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| Fiction.LiveBench | — | 97.2% |
| LMArena Longer Query | 1301 | — |
Writing & Preference o3-pro leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), o3-pro: 57.1 (#133)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | o3-pro |
|---|---|---|
| LMArena Text | 1294 | — |
| LMArena Creative Writing | 1285 | — |
| Short-Story Creative Writing | — | 84.4% |
| LMArena Multi-Turn | 1297 | — |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than o3-pro?
o3-pro is the stronger model overall, scoring 42.9 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or o3-pro better for coding?
o3-pro scores higher on coding benchmarks: 55.5 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and o3-pro share?
1 benchmark has published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and o3-pro has 12.