Model comparison
DeepSeek-V2.5 (Sep 2024) vs Nemotron 3 Ultra
Nemotron 3 Ultra is the stronger model overall, scoring 42.5 to 37.6 on the Noometry Index.
Last verified . 16 shared benchmarks.
DeepSeek
NVIDIASummary
- They share 16 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 1 category and Nemotron 3 Ultra in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Nemotron 3 Ultra leads 52.5 to 34.8.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra | |
|---|---|---|
| Provider | DeepSeek | NVIDIA |
| Noometry Index | 37.6 | 42.5 |
| Released | 2024-09-06 | 2026-06-04 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 128K |
| Input $ / M tokens | — | $0.50 |
| Output $ / M tokens | — | $2.20 |
| Results tracked | 22 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Nemotron 3 Ultra: 38.1 (#182)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Coding | 1309 | 1468 |
| FrontierCode | — | 13.6% |
| Aider Polyglot | 17.8% | — |
| SciCode | — | 40.3% |
| WeirdML | — | 43.5% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Nemotron 3 Ultra: 23.4 (#126)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| APEX-Agents | — | 22.7% |
Reasoning Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Nemotron 3 Ultra: 31.0 (#82)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1452 |
| CritPt | — | 3.1% |
| Chess Puzzles | — | 12% |
| Mystery Game Puzzles | — | 20% |
| DTBench | — | 90.1% |
| LMCA | — | 36.9% |
| Epoch Capabilities Index | — | 146.17 |
Math Too close to call
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Nemotron 3 Ultra: 35.3 (#187)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Math | 1288 | 1457 |
| OTIS Mock AIME 2024-2025 | — | 86.7% |
| ProofBench | — | 2% |
Knowledge Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Nemotron 3 Ultra: 52.5 (#61)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Expert | 1266 | 1472 |
| GPQA Diamond | — | 85.4% |
Multilingual Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Nemotron 3 Ultra: 53.5 (#64)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Non-English | 1273 | 1427 |
| LMArena Chinese | 1318 | 1497 |
| LMArena French | 1289 | 1472 |
| LMArena German | 1258 | 1471 |
| LMArena Korean | 1209 | 1386 |
| LMArena Russian | 1289 | 1417 |
| LMArena Spanish | 1248 | 1454 |
| LMArena Japanese | 1228 | — |
Instruction Following Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Nemotron 3 Ultra: 74.7 (#88)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Instruction Following | 1280 | 1418 |
Long Context Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Nemotron 3 Ultra: 43.9 (#81)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Longer Query | 1301 | 1437 |
Writing & Preference Nemotron 3 Ultra leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Nemotron 3 Ultra: 66.3 (#36)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Nemotron 3 Ultra |
|---|---|---|
| LMArena Text | 1294 | 1445 |
| LMArena Creative Writing | 1285 | 1398 |
| LMArena Multi-Turn | 1297 | 1411 |
| EQ-Bench Creative Writing | — | 1692 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Nemotron 3 Ultra?
Nemotron 3 Ultra is the stronger model overall, scoring 42.5 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Nemotron 3 Ultra better for coding?
Nemotron 3 Ultra scores higher on coding benchmarks: 38.1 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Nemotron 3 Ultra share?
16 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Nemotron 3 Ultra has 30.