NVIDIA · Nemotron 3
Nemotron 3 Ultra 550B A55B
NVIDIA's open-weight (weights, data and recipes) hybrid Mamba-Transformer LatentMoE flagship for agentic, reasoning and long-context workloads.
- AA Intelligence
- 22.9#61 of 142
- Input / output per 1M
- $0.60/ $2.50
- Output speed
- 140tok/s
- Context window
- 1M tokens
Composite indexes
- Artificial Analysis Intelligence Index#61 of 14223
Leader: Claude Opus 5.5 at 58
Coding
- SciCode#68 of 9444.6%40.3%
Leader: Claude Opus 5.5 at 66.9%
- Terminal-Bench 4.0#55 of 880.5%
Leader: Claude Mythos 5.1 at 60.9%
- Terminal-Bench 2.x#57 of 10356.4%53.9%
Leader: Claude Fable 5.1 at 91.4%
- SWE-bench Verified#9 of 1670.7%
Leader: Claude Sonnet 5 at 85.2%
- LiveCodeBench#4 of 5689%
Leader: Fugu at 92.9%
Agents and tool use
- τ²-Bench#30 of 8883.3%
Leader: GLM-5.2 at 99.1%
- GDPval-AA#52 of 1031000
Leader: Claude Opus 5.5 at 1846
- BrowseComp#11 of 1444.4%
Leader: GPT-6 Astra at 91.5%
Reasoning and knowledge
- GPQA Diamond#51 of 14287%86.7%
Leader: GPT-6 Astra at 96.1%
- Humanity’s Last Exam#58 of 14737.4%28.4%26.7%
Leader: Claude Opus 5.5 at 61.4%
- CritPt#57 of 1393.1%3.1%
Leader: GPT-5.6 Sol at 32.3%
- MMLU-Pro#1 of 1486.8%
Leader: Nemotron 3 Ultra 550B A55B at 86.8%
Vision and long context
- AA Long Context Reasoning#40 of 13979.3%65.4%
Leader: Kimi K3 at 88.7%