Model comparison
Nemotron 3 Ultra vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 42.5 on the Noometry Index. Nemotron 3 Ultra costs 3.2× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 27 shared benchmarks.
NVIDIA
Alibaba (Qwen)Summary
- They share 27 benchmarks with published results for both. Nemotron 3 Ultra scores higher in 0 categories and Qwen3.8 Max in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 35.3.
- The biggest single-benchmark swing is ProofBench: 2% for Nemotron 3 Ultra and 58% for Qwen3.8 Max.
- Nemotron 3 Ultra is cheaper at $0.50 / $2.20 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- Qwen3.8 Max accepts more context: 1M tokens versus 262K.
- Nemotron 3 Ultra has downloadable open weights; the other is API-only.
Side by side
| Nemotron 3 Ultra | Qwen3.8 Max | |
|---|---|---|
| Provider | NVIDIA | Alibaba (Qwen) |
| Noometry Index | 42.5 | 56.8 |
| Released | 2026-06-04 | 2026-08-02 |
| Weights | Open | Proprietary |
| Context window | 262K | 1M |
| Max output | 128K | 131K |
| Input $ / M tokens | $0.50 | $2 |
| Output $ / M tokens | $2.20 | $6 |
| Results tracked | 30 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 Max leads
Nemotron 3 Ultra: 38.1 (#182), Qwen3.8 Max: 53.5 (#29)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| SciCode | 40.3% | 53.2% |
| LMArena Coding | 1468 | 1502 |
| DeepSWE | — | 57.5% |
| FrontierCode | 13.6% | — |
| LMArena WebDev | — | 1674 |
| FrontierSWE | — | 17.8% |
| WeirdML | 43.5% | — |
Agentic & Tool Use Qwen3.8 Max leads
Nemotron 3 Ultra: 23.4 (#126), Qwen3.8 Max: 45.4 (#14)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | 22.7% | 63.3% |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |
Reasoning Qwen3.8 Max leads
Nemotron 3 Ultra: 31.0 (#82), Qwen3.8 Max: 54.4 (#26)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| CritPt | 3.1% | 20% |
| Chess Puzzles | 12% | 40% |
| LMArena Hard Prompts | 1452 | 1496 |
| Mystery Game Puzzles | 20% | 38% |
| DTBench | 90.1% | 92% |
| LMCA | 36.9% | 46.2% |
| Epoch Capabilities Index | 146.17 | 156.41 |
| NYT Connections (extended) | — | 88.3% |
Math Qwen3.8 Max leads
Nemotron 3 Ultra: 35.3 (#187), Qwen3.8 Max: 73.2 (#20)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 86.7% | 100% |
| ProofBench | 2% | 58% |
| LMArena Math | 1457 | 1499 |
| FrontierMath (Tiers 1-3) | — | 74.7% |
| FrontierMath Tier 4 | — | 46.3% |
Knowledge Qwen3.8 Max leads
Nemotron 3 Ultra: 52.5 (#61), Qwen3.8 Max: 61.7 (#27)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 85.4% | 92.7% |
| LMArena Expert | 1472 | 1507 |
| SimpleQA Verified | — | 47.3% |
Multimodal Not comparable
Nemotron 3 Ultra: —, Qwen3.8 Max: 37.2 (#75)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | — | 1314 |
| Furniture Assembly | — | 20% |
Multilingual Qwen3.8 Max leads
Nemotron 3 Ultra: 53.5 (#64), Qwen3.8 Max: 56.7 (#18)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1427 | 1472 |
| LMArena Chinese | 1497 | 1538 |
| LMArena French | 1472 | 1503 |
| LMArena German | 1471 | 1483 |
| LMArena Korean | 1386 | 1461 |
| LMArena Russian | 1417 | 1481 |
| LMArena Spanish | 1454 | 1492 |
| LMArena Japanese | — | 1467 |
Instruction Following Qwen3.8 Max leads
Nemotron 3 Ultra: 74.7 (#88), Qwen3.8 Max: 77.6 (#17)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1418 | 1479 |
Long Context Qwen3.8 Max leads
Nemotron 3 Ultra: 43.9 (#81), Qwen3.8 Max: 45.6 (#31)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1437 | 1489 |
Writing & Preference Too close to call
Nemotron 3 Ultra: 66.3 (#36), Qwen3.8 Max: 67.1 (#30)
| Benchmark | Nemotron 3 Ultra | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1445 | 1483 |
| LMArena Creative Writing | 1398 | 1479 |
| LMArena Multi-Turn | 1411 | 1489 |
| EQ-Bench Creative Writing | 1692 | — |
Frequently asked questions
Is Nemotron 3 Ultra better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 42.5 on the Noometry Index. Nemotron 3 Ultra costs 3.2× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, Nemotron 3 Ultra or Qwen3.8 Max?
Nemotron 3 Ultra is cheaper. It lists at $0.50 per million input tokens and $2.20 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is Nemotron 3 Ultra or Qwen3.8 Max better for coding?
Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 38.1 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 Max does, with 1M tokens against 262K.
How many benchmarks do Nemotron 3 Ultra and Qwen3.8 Max share?
27 benchmarks have published results for both models. Nemotron 3 Ultra has 30 scored results on Noometry and Qwen3.8 Max has 39.