Nemotron 3 Nano 30B A3B
Nemotron 3 Nano 30B A3BRankingin thinking mode?
Alternatives?
CheaperGPT-5.6 Lunamedium×2.8-7%$0.016BetterGPT-6 Lunalow×2.4-73%$0.005Function calling only with none · Chat Completions+3
OpenQwen3.8 27Bxhigh×3.8+5,914%$1.01max_tokens range · QwenCloud+4
CheaperLing 3.0 Flash VLthinkingOpen×2.8-14%$0.06 / $0.18Memory (4-bit): 19 GB → 75 GBCheaperLing 3.0 Flash FinthinkingOpen×2.5-14%$0.06 / $0.18Memory (4-bit): 19 GB → 76 GBGPT-5.6 Lunamedium×2.8-7%$0.016Ling 3.0 Flash VLthinkingOpen×2.8-14%$0.06 / $0.18Memory (4-bit): 19 GB → 75 GBLing 3.0 Flash FinthinkingOpen×2.5-14%$0.06 / $0.18Memory (4-bit): 19 GB → 76 GBGPT-6 Lunalow×2.4-73%$0.005Function calling only with none · Chat Completions+3Ling 3.0 FlashthinkingOpen×2.3-14%$0.06 / $0.18Memory (4-bit): 19 GB → 76 GBGPT-6 Lunalow×2.4-73%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.4-41%$0.01
Mistral Small 4thinkingOpen+27%-10%$0.015Memory (4-bit): 19 GB → 72 GB
MiMo v2.5thinkingOpen×2.5+14%$0.019Memory (4-bit): 19 GB → 186 GB
Granite 4.2 8BthinkingOpen+25%+41%$0.024Context: 131K → 128KQwen3.8 27BxhighOpen×3.8+5,914%$1.01max_tokens range · QwenCloud+4IK2 Horizon MoVA 36B A4BthinkingOpen×2.8+1,567%$1 / $4Memory (4-bit): 19 GB → 22 GBQwen 3.5 27BthinkingOpen×2.6+686%$0.30 / $2.40AG9v3 39A5BOpen×2.5——Memory (4-bit): 19 GB → 23 GBQwen3.6 27BthinkingOpen×2.4+3,609%$0.62Nemotron 3.5 Super VLOpen×3+1,024%$0.19Memory (4-bit): 19 GB → 75 GBNemotron 3 UltrathinkingOpen×2.6+3,344%$0.58Memory (4-bit): 19 GB → 336 GBNemotron 3.5 Bolt 74B-A9BOpen×2.4——Memory (4-bit): 19 GB → 45 GBNemotron 3.5 LightningthinkingOpen+45%+375%$0.079Memory (4-bit): 19 GB → 20 GBNemotron 3 Super 120B A12BOpen+44%+9,685%$1.64Memory (4-bit): 19 GB → 74 GBNemotron Cascade 2 30B A3BOpen+31%——Nemotron 3 Nano Omni 30B A3B ReasoningOpen+15%—$0 / $0Memory (4-bit): 19 GB → 20 GBLlama Nemotron Super 49B v1.5non-thinkingOpen-17%—$0.40 / $0.40Context: 131K → 128K+1All NVIDIA models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Nemotron 3 Nano 30B A3B cost?›
$0.06 per 1M input tokens and $0.24 per 1M output tokens (cheapest host: OpenRouter). A typical benchmark task costs about $0.017 at default effort.
Is Nemotron 3 Nano 30B A3B free to use?›
Yes, there is a free API option.
How good is Nemotron 3 Nano 30B A3B, and what is it best at?›
#168 of 334 in Chat & Knowledge, #75 of 132 in Coding, #110 of 133 in Agents & Tool Use, #163 of 195 in Writing & Language, #119 of 133 in Vision & Documents. Strongest at: Instructions.
Is Nemotron 3 Nano 30B A3B better than Gemma 4 E4B?›
On AA Intelligence Index 4.3.2, Nemotron 3 Nano 30B A3B scores 8.9 vs 8.9 — about the same. It is 105% more expensive per task.
How can I use Nemotron 3 Nano 30B A3B?›
Through the API of DeepInfra, Nebius, Novita.
What is the context window of Nemotron 3 Nano 30B A3B?›
131K tokens, with up to 131K output tokens.
When was Nemotron 3 Nano 30B A3B released? Is there a newer version?›
Released 2025-12-04. It is the current version in its line.
Is Nemotron 3 Nano 30B A3B open source?›
Yes, the weights are open (nvidia-nemotron-open-model-license licence).
Can I run Nemotron 3 Nano 30B A3B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Nemotron 3 Nano 30B A3B read images?›
No, it accepts text only.
What is Nemotron 3 Nano 30B A3B?›
Nemotron 3 Nano 30B A3B is an open-weights language model from NVIDIA, released Dec 2025.
What are the best alternatives to Nemotron 3 Nano 30B A3B?›
GPT-5.6 Luna (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).