NVIDIA Nemotron 49B v1 Card - GPQA: leaderboard
GPQA graduate-level science multiple-choice questions, pass@1 with reasoning on and off, using a step-by-step prompt that asks for a boxed answer letter. From NVIDIA's own Llama-3.3-Nemotron-Super-49B-v1 model card, which states a 32k sequence length with benchmarks run up to 16 times and averaged, and recommends temperature 0.6 with top-p 0.95 for reasoning on and greedy decoding for reasoning off.
Metric: pass@1. Source: huggingface.co. Saturation forecast: Around February 2027. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 Nemotron Super 49B (v1) (Reasoning) | 66.67 | #1336 (Llama 3.3 Nemotron Super 49B (v1)) |
| 2 | Llama 3.3 Nemotron Super 49B (v1) (Non-reasoning) | 50 | #1336 (Llama 3.3 Nemotron Super 49B (v1)) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=nvidia-nemotron-49b-v1-card-gpqa · How It Works · Data refreshed daily, snapshot 2026-11-12.