Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | DeepSeek V3.1 (Non-reasoning) | Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) |
|---|---|---|
| Input ($/M tokens) | $0.56 | $0.6 |
| Output ($/M tokens) | $1.68 | $1.8 |
Data from Artificial Analysis API — 12 benchmarks
DeepSeek V3.1 (Non-reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.84/M tokens vs $0.90/M for Llama 3.1 Nemotron Ultra 253B v1 (Reasoning).
It's a tie — both models win 6 benchmarks each across 12 evaluated categories. See the detailed benchmark chart above for per-category results.
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) generates tokens faster at 42 tok/s vs 0 tok/s. DeepSeek V3.1 (Non-reasoning) also has lower time-to-first-token (0.00s vs 0.68s).
Choose based on your priorities: DeepSeek V3.1 (Non-reasoning) for lower cost, both perform similarly on benchmarks, and Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.