Model comparison
Claude Sonnet 4 vs Nemotron 3 Ultra
Nemotron 3 Ultra is the stronger model overall, scoring 44.7 to 40.6 on the Noometry Index.
Last verified . 25 shared benchmarks.
NVIDIASummary
- They share 25 benchmarks with published results for both. Claude Sonnet 4 scores higher in 2 categories and Nemotron 3 Ultra in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Nemotron 3 Ultra leads 52.5 to 41.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 71.1% for Claude Sonnet 4 and 86.7% for Nemotron 3 Ultra.
- Nemotron 3 Ultra is cheaper at $0.60 / $2.40 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Nemotron 3 Ultra accepts more context: 262K tokens versus 200K.
- Nemotron 3 Ultra has downloadable open weights; the other is API-only.
Side by side
| Claude Sonnet 4 | Nemotron 3 Ultra | |
|---|---|---|
| Provider | Anthropic | NVIDIA |
| Noometry Index | 40.6 | 44.7 |
| Released | 2025-05-22 | 2026-06-04 |
| Weights | Proprietary | Open |
| Context window | 200K | 262K |
| Max output | 64K | 128K |
| Input $ / M tokens | $3 | $0.60 |
| Output $ / M tokens | $15 | $2.40 |
| Results tracked | 58 | 29 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude Sonnet 4: 44.0 (#70), Nemotron 3 Ultra: 37.7 (#174)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| SciCode | 40% | 40.3% |
| WeirdML | 46.1% | 43.5% |
| LMArena Coding | 1414 | 1468 |
| FrontierCode | — | 13.6% |
| SWE-bench Verified (bash only) | 64.9% | — |
| Aider Polyglot | 61.3% | — |
| GSO | 4.9% | — |
| ALE-Bench | 655.35 | — |
Agentic & Tool Use Not comparable
Claude Sonnet 4: 38.3 (#29), Nemotron 3 Ultra: —
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| TheAgentCompany | 33.1% | — |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| METR Time Horizons | 62% | — |
Reasoning Nemotron 3 Ultra leads
Claude Sonnet 4: 23.0 (#169), Nemotron 3 Ultra: 30.3 (#78)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| CritPt | 0.3% | 3.1% |
| LMArena Hard Prompts | 1372 | 1452 |
| DTBench | 77.1% | 90.1% |
| LMCA | 29% | 36.9% |
| Epoch Capabilities Index | 141.69 | 146.2 |
| ARC-AGI-2 | 5.9% | — |
| SimpleBench | 45.5% | — |
| Kagi LLM Benchmark | 73% | — |
| ARC-AGI-1 | 40% | — |
| Chess Puzzles | — | 12% |
| EnigmaEval | 3.1% | — |
| Mystery Game Puzzles | — | 20% |
| ForecastBench | 60.2 | — |
Math Claude Sonnet 4 leads
Claude Sonnet 4: 41.5 (#71), Nemotron 3 Ultra: 34.4 (#160)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 86.7% |
| LMArena Math | 1375 | 1456 |
| ProofBench | — | 2% |
| Omni-MATH | 60.2% | — |
| MATH Level 5 | 84.4% | — |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Nemotron 3 Ultra leads
Claude Sonnet 4: 41.6 (#99), Nemotron 3 Ultra: 52.5 (#56)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| GPQA Diamond | 79.2% | 85.4% |
| LMArena Expert | 1371 | 1472 |
| Humanity's Last Exam | 7.8% | — |
| MMLU-Pro | 84.3% | — |
| Confabulations | 13.2% | — |
| Vectara Hallucination Rate | 10.3% | — |
| GPQA (HELM) | 70.6% | — |
Multimodal Not comparable
Claude Sonnet 4: 25.6 (#116), Nemotron 3 Ultra: —
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Vision | 1191 | — |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| MindCube | 44.8% | — |
Multilingual Nemotron 3 Ultra leads
Claude Sonnet 4: 46.9 (#151), Nemotron 3 Ultra: 53.7 (#61)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Non-English | 1333 | 1427 |
| LMArena Chinese | 1351 | 1498 |
| LMArena French | 1364 | 1475 |
| LMArena German | 1331 | 1471 |
| LMArena Korean | 1293 | 1387 |
| LMArena Russian | 1355 | 1415 |
| LMArena Spanish | 1356 | 1452 |
| LMArena Japanese | 1301 | — |
Instruction Following Nemotron 3 Ultra leads
Claude Sonnet 4: 71.8 (#141), Nemotron 3 Ultra: 74.9 (#84)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Instruction Following | 1375 | 1417 |
| IFEval | 84% | — |
Long Context Nemotron 3 Ultra leads
Claude Sonnet 4: 33.6 (#254), Nemotron 3 Ultra: 43.9 (#77)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Longer Query | 1398 | 1437 |
| Fiction.LiveBench | 46.9% | — |
Writing & Preference Nemotron 3 Ultra leads
Claude Sonnet 4: 57.3 (#127), Nemotron 3 Ultra: 66.3 (#35)
| Benchmark | Claude Sonnet 4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Text | 1351 | 1444 |
| LMArena Creative Writing | 1345 | 1392 |
| EQ-Bench Creative Writing | 1483 | 1692 |
| LMArena Multi-Turn | 1376 | 1412 |
| Short-Story Creative Writing | 81.4% | — |
| WildBench | 83.8% | — |
Frequently asked questions
Is Claude Sonnet 4 better than Nemotron 3 Ultra?
Nemotron 3 Ultra is the stronger model overall, scoring 44.7 to 40.6 on the Noometry Index.
Which is cheaper, Claude Sonnet 4 or Nemotron 3 Ultra?
Nemotron 3 Ultra is cheaper. It lists at $0.60 per million input tokens and $2.40 per million output tokens; Claude Sonnet 4 lists at $3 and $15.
Is Claude Sonnet 4 or Nemotron 3 Ultra better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 44.0 versus 37.7 in the Noometry coding category.
Which has the bigger context window?
Nemotron 3 Ultra does, with 262K tokens against 200K.
How many benchmarks do Claude Sonnet 4 and Nemotron 3 Ultra share?
25 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and Nemotron 3 Ultra has 29.