Nemotron 3 Ultra
Nemotron 3 UltraAlternatives?
CheaperCommand A+thinkingOpen×2.1-43%$0.30 / $1.50Context: 1M → 128K
BetterGranite 4.2 3BthinkingOpen+13%-98%$0.006Context: 1M → 128KOOpenMiniCPM5 1Bnon-thinking×33——Context: 1M → 131KlCheaperLFM2.5 2.6BthinkingOpen+86%-63%$0.35 / $0.50Context: 1M → 131KMCheaperQuasar 438B+38%-60%$0.15Command A+thinkingOpen×2.1-43%$0.30 / $1.50Context: 1M → 128KlLFM2.5 2.6BthinkingOpen+86%-63%$0.35 / $0.50Context: 1M → 131KMQuasar 438B+38%-60%$0.15
Ling 3.0 Flash VLthinkingOpen+35%-91%$0.06 / $0.18Context: 1M → 262K
MiMo v2.5 ProthinkingOpen+20%-59%$0.15Memory (4-bit): 336 GB → 614 GBGranite 4.2 3BthinkingOpen+13%-98%$0.006Context: 1M → 128KGranite 4.2 30BthinkingOpen+16%-80%$0.072Context: 1M → 128KMQuasar 438B+38%-60%$0.15MiMo v2.5 ProthinkingOpen+20%-59%$0.15Memory (4-bit): 336 GB → 614 GB
Grok 4.3high+15%-44%$0.21Command A+thinkingOpen×2.1-43%$0.30 / $1.50Context: 1M → 128KlLFM2.5 2.6BthinkingOpen+86%-63%$0.35 / $0.50Context: 1M → 131KNVIDIA Nemotron Nano 12B v2 VLnon-thinkingOpen-69%—$0.20 / $0.60Context: 1M → 128KNemotron 3 Nano 30B A3Bnon-thinkingOpen-67%—$0.06 / $0.24Context: 1M → 131KNVIDIA Nemotron 3 Nano 4BthinkingOpen-66%——Context: 1M → 262KNemotron 3 Super 120B A12BOpen-66%+347%$1.64Nemotron 3 Nano Omni 30B A3B ReasoningOpen-65%—$0 / $0Context: 1M → 256KNemotron Cascade 2 30B A3BOpen-65%——NVIDIA Nemotron Nano 9B V2non-thinkingOpen-60%—$0.04 / $0.16Context: 1M → 131KLlama Nemotron Super 49B v1.5non-thinkingOpen-56%—$0.40 / $0.40Context: 1M → 128KAll NVIDIA models →Run it yourself
estimateBenchmarks
in Chat & KnowledgeFAQ
How much does Nemotron 3 Ultra cost?›
$0.60 per 1M input tokens and $2.40 per 1M output tokens (cheapest host: OpenRouter). A typical benchmark task costs about $0.37 at default effort.
Is Nemotron 3 Ultra free to use?›
Yes, there is a free API option.
How good is Nemotron 3 Ultra, and what is it best at?›
#92 of 336 in Chat & Knowledge, #65 of 134 in Coding, #94 of 135 in Agents & Tool Use, #3 of 245 in Writing & Language, #89 of 135 in Vision & Documents. Strongest at: Science, Long context, Knowledge, Instructions.
Is Nemotron 3 Ultra better than GLM 5.3?›
On AA Omniscience · Hallucination Rate, Nemotron 3 Ultra scores 29.7 vs 29.6 — about the same. It is 82% cheaper per task.
How can I use Nemotron 3 Ultra?›
Through the API of BaseTen, DeepInfra, Venice.
What is the context window of Nemotron 3 Ultra?›
1M tokens, with up to 66K output tokens.
When was Nemotron 3 Ultra released? Is there a newer version?›
Released 2026-06-04. It is the current version in its line.
Is Nemotron 3 Ultra open source?›
Yes, the weights are open (openmdw-1.1 licence).
Can I run Nemotron 3 Ultra locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Nemotron 3 Ultra read images?›
No, it accepts text only.
What is Nemotron 3 Ultra?›
Nemotron 3 Ultra is an open-weights language model from NVIDIA, released Jun 2026.
What are the best alternatives to Nemotron 3 Ultra?›
Command A+ (cheaper), Granite 4.2 3B (better), MiniCPM5 1B (open weights).