gpt-oss-120b is 27.8× cheaper per million tokens (blended). Nemotron 3 Ultra has the larger context window (202,800 vs 131,072 tokens, 1.5×).
Nemotron 3 Ultra vs gpt-oss-120b
Live catalog fields. Best-in-row is highlighted.
Nemotron 3 Ultraopen
NVIDIA · nvidia__nemotron-3-ultra-550b-a55b
gpt-oss-120bopen
OpenAI · openai__gpt-oss-120b
Identity
| Field | nvidia__nemotron-3-ultra-550b-a55b | openai__gpt-oss-120b |
|---|---|---|
| Provider | NVIDIA | OpenAI |
| Slug | nvidia__nemotron-3-ultra-550b-a55b | openai__gpt-oss-120b |
| Status | live | live |
| Open weights | Yes | Yes |
| License | — | — |
| HuggingFace | nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | openai/gpt-oss-120b |
Quality
| Field | nvidia__nemotron-3-ultra-550b-a55b | openai__gpt-oss-120b |
|---|---|---|
| Quality index (0–100) | — | — |
Cost
| Field | nvidia__nemotron-3-ultra-550b-a55b | openai__gpt-oss-120b |
|---|---|---|
| Input $ / 1M tokens | $0.60 | $0.02 |
| Cached input $ / 1M | $0.12 | $0.02 |
| Output $ / 1M tokens | $2.40 | $0.09 |
Context
| Field | nvidia__nemotron-3-ultra-550b-a55b | openai__gpt-oss-120b |
|---|---|---|
| Max context tokens | 202,800 | 131,072 |
| Max output tokens | 182,520 | 65,536 |
Modalities
| Field | nvidia__nemotron-3-ultra-550b-a55b | openai__gpt-oss-120b |
|---|---|---|
| Input | text | text |
| Output | text | text |