Model comparison
GPT-5.4 mini vs Nemotron 3 Ultra
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 42.5 on the Noometry Index. Nemotron 3 Ultra costs 1.6× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Last verified . 29 shared benchmarks.
NVIDIASummary
- They share 29 benchmarks with published results for both. GPT-5.4 mini scores higher in 3 categories and Nemotron 3 Ultra in 6 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 mini leads 45.5 to 35.3.
- The biggest single-benchmark swing is ProofBench: 21% for GPT-5.4 mini and 2% for Nemotron 3 Ultra.
- Nemotron 3 Ultra is cheaper at $0.60 / $2.40 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
- GPT-5.4 mini accepts more context: 400K tokens versus 262K.
- Nemotron 3 Ultra has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 mini | Nemotron 3 Ultra | |
|---|---|---|
| Provider | OpenAI | NVIDIA |
| Noometry Index | 45.0 | 42.5 |
| Released | 2026-03-17 | 2026-06-04 |
| Weights | Proprietary | Open |
| Context window | 400K | 262K |
| Max output | 128K | 128K |
| Input $ / M tokens | $0.75 | $0.60 |
| Output $ / M tokens | $4.50 | $2.40 |
| Results tracked | 46 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 mini leads
GPT-5.4 mini: 45.2 (#72), Nemotron 3 Ultra: 38.1 (#182)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| FrontierCode | 27% | 13.6% |
| SciCode | 49.9% | 40.3% |
| WeirdML | 60.3% | 43.5% |
| LMArena Coding | 1438 | 1468 |
| LMArena WebDev | 1397 | — |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use GPT-5.4 mini leads
GPT-5.4 mini: 29.9 (#81), Nemotron 3 Ultra: 23.4 (#126)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| APEX-Agents | — | 22.7% |
| DeepResearch Bench | 36.3% | — |
Reasoning Too close to call
GPT-5.4 mini: 30.4 (#85), Nemotron 3 Ultra: 31.0 (#82)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| CritPt | 10% | 3.1% |
| Chess Puzzles | 24% | 12% |
| LMArena Hard Prompts | 1424 | 1452 |
| Mystery Game Puzzles | 11% | 20% |
| DTBench | 80% | 90.1% |
| LMCA | 40.8% | 36.9% |
| Epoch Capabilities Index | 148.84 | 146.17 |
| ARC-AGI-2 | 18.9% | — |
| Kagi LLM Benchmark | 37.9% | — |
| NYT Connections (extended) | 61.8% | — |
| ARC-AGI-1 | 63.7% | — |
| Thematic Generalization | 61.7% | — |
| ForecastBench | 57 | — |
Math GPT-5.4 mini leads
GPT-5.4 mini: 45.5 (#75), Nemotron 3 Ultra: 35.3 (#187)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 86.7% |
| ProofBench | 21% | 2% |
| LMArena Math | 1419 | 1457 |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 9.8% | — |
| FrontierMath (Feb 2025 set) | 28.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge Nemotron 3 Ultra leads
GPT-5.4 mini: 51.5 (#67), Nemotron 3 Ultra: 52.5 (#61)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| GPQA Diamond | 86.9% | 85.4% |
| LMArena Expert | 1435 | 1472 |
| SimpleQA Verified | 29.4% | — |
| Vectara Hallucination Rate | 5.5% | — |
Multimodal Not comparable
GPT-5.4 mini: 39.7 (#56), Nemotron 3 Ultra: —
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| LMArena Vision | 1245 | — |
Multilingual Nemotron 3 Ultra leads
GPT-5.4 mini: 51.9 (#96), Nemotron 3 Ultra: 53.5 (#64)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| LMArena Non-English | 1405 | 1427 |
| LMArena Chinese | 1446 | 1497 |
| LMArena French | 1440 | 1472 |
| LMArena German | 1409 | 1471 |
| LMArena Korean | 1368 | 1386 |
| LMArena Russian | 1417 | 1417 |
| LMArena Spanish | 1405 | 1454 |
| LMArena Japanese | 1374 | — |
Instruction Following Too close to call
GPT-5.4 mini: 74.1 (#102), Nemotron 3 Ultra: 74.7 (#88)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| LMArena Instruction Following | 1405 | 1418 |
Long Context Too close to call
GPT-5.4 mini: 43.0 (#112), Nemotron 3 Ultra: 43.9 (#81)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| LMArena Longer Query | 1407 | 1437 |
Writing & Preference Nemotron 3 Ultra leads
GPT-5.4 mini: 64.0 (#58), Nemotron 3 Ultra: 66.3 (#36)
| Benchmark | GPT-5.4 mini | Nemotron 3 Ultra |
|---|---|---|
| LMArena Text | 1412 | 1445 |
| LMArena Creative Writing | 1370 | 1398 |
| EQ-Bench Creative Writing | 1665 | 1692 |
| LMArena Multi-Turn | 1429 | 1411 |
Frequently asked questions
Is GPT-5.4 mini better than Nemotron 3 Ultra?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 42.5 on the Noometry Index. Nemotron 3 Ultra costs 1.6× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 mini 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; GPT-5.4 mini lists at $0.75 and $4.50.
Is GPT-5.4 mini or Nemotron 3 Ultra better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 38.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 mini does, with 400K tokens against 262K.
How many benchmarks do GPT-5.4 mini and Nemotron 3 Ultra share?
29 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Nemotron 3 Ultra has 30.