Nemotron 4: pricing, context window and benchmarks
Nemotron 4 is a multimodal model from NVIDIA. It offers a 65,536-token context window, costs $4.18 per million input tokens and $18.05 per million output tokens, and averages 76.6 across the index's benchmark suite.
Pricing
Nemotron 4 costs $4.18 per million input tokens and $18.05 per million output tokens. A blended 3:1 input-to-output mix works out to about $7.65 per million tokens. Cached input is listed at $0.58 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $4.18 | per 1M tokens |
| Output | $18.05 | per 1M tokens |
| Cached input | $0.58 | per 1M tokens |
| Blended 3:1 | $7.65 | estimate per 1M tokens |
| 1M in + 1M out | $22.23 | estimate |
Benchmarks
Scores below are modelled estimates on a 0–100 scale. Use them to compare relative position, not as official vendor results.
| Benchmark | Score | Rank | Percentile | Relative |
|---|---|---|---|---|
| MMLU-Pro | 77.8 | 250 | 58th | |
| GPQA Diamond | 74.4 | 158 | 74th | |
| SWE-bench Verified | 63.7 | 240 | 60th | |
| HumanEval | 82.7 | 164 | 73th | |
| MATH-500 | 77.7 | 262 | 56th | |
| AIME 2025 | 72.0 | 221 | 63th | |
| LiveCodeBench | 70.1 | 191 | 68th | |
| IFEval | 82.2 | 270 | 55th | |
| TruthfulQA | 80.1 | 176 | 71th | |
| HellaSwag | 86.1 | 214 | 64th | |
| MMMU | 76.1 | 136 | 77th | |
| Arena Elo | 1409.7 | 184 | 69th |
What Nemotron 4 is good for
Nemotron 4 is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $4.18 per million input tokens, it sits in the mid-range of the index, and the 65,536-token context window is enough for large files and multi-file diffs.
Strengths
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Accepts more than text, including image input where the provider supports it.
- Coding benchmarks are a relative strength within its tier.
Trade-offs
- No standout weakness in the indexed specifications, but real-world behaviour still needs testing.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Llama 3.3 Instruct | Meta AI | 128000 | $4.33 | 76.6 |
| Nemotron 4 | NVIDIA | 65536 | $4.18 | 76.6 |
| Qwen3-Coder Base | Alibaba Cloud | 200000 | $3.99 | 76.6 |
| Claude Sonnet 4.5 Mini | Anthropic | 32768 | $11.43 | 76.5 |
| Qwen3-Max Mini | Alibaba Cloud | 128000 | $11.13 | 76.5 |
| Yi-Lightning Instruct | 01.AI | 128000 | $4.16 | 76.5 |
FAQ
How much does Nemotron 4 cost?
Nemotron 4 is listed at $4.18 per million input tokens and $18.05 per million output tokens. Cached input is $0.58 per million tokens. A blended 3:1 input-to-output mix is about $7.65 per million tokens.
What is the context window of Nemotron 4?
Nemotron 4 has a context window of 65,536 tokens and can emit up to 32,768 output tokens per request in this index.
Is Nemotron 4 open weights?
Yes. Nemotron 4 is listed as an open-weight model under a open weights licence.
How does Nemotron 4 score on benchmarks?
Nemotron 4 averages 76.6 across the scored benchmarks in this index, ranking #204 of 600 models. Its strongest relative result is Arena Elo at 1409.7.
What is Nemotron 4 best for?
Nemotron 4 is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $4.18 per million input tokens, it sits in the mid-range of the index, and the 65,536-token context window is enough for large files and multi-file diffs.