Nemotron 3.5 Lightning 30B API
Model source: Hugging FaceReasons and uses tools while spending very little, built to answer fast.
Nemotron 3.5 Lightning is NVIDIA's mixture-of-experts model built to answer fast: 30 billion parameters, 3 billion active per token, so it reasons and calls tools while spending very little. On GPU Flow it costs €0.12 per million input tokens and €0.30 per million output, with a 131,072-token context.
Reads in one go
≈ 280 pages
131,072 tokens
Brain size
Mid-sized
30B MoE, 3B active, NVFP4
Where it runs
Spain (EU)
MIG B200
Understands
Text, not images
Reasoning, Tools
What it is good at
- Multi-step reasoning at speed. Works through a problem in several steps without the wait of a large model.
- Tool calling for agents that run many times. Cheap per call, so an agent can loop through tools again and again.
- Long inputs. A context of 131,072 tokens, about 280 pages.
Text only.
Pricing
€ per million tokens, VAT included
For your technical team
Compatible with the OpenAI SDK: change the base URL and the key, and that is it.
curl https://api.gpuflow.ai/v1/chat/completions \
-H "Authorization: Bearer $GPUFLOW_API_KEY" \
-d '{ "model": "nemotron-3.5-lightning-30b-a3b",
"messages": [{ "role": "user", "content": "Hola" }] }'Context and limits
- Context
- 131,072 tokens per request, about 280 pages.
- Requests per minute
- 1,000 per key, with bursts up to double; raisable to 5,000 from the dashboard. 5,000 in total per account.
- Tokens per minute
- 2,000,000 input (cached tokens do not count) and 400,000 output, per key.
- Prompt per request
- Up to 256,000 tokens, raisable on request.
- If you go over
- A 429 error with a Retry-After header telling you how many seconds to wait.
Nemotron 3.5 Lightning 30B compared with similar models
| Nemotron 3.5 Lightning 30B | Qwen3.8 27B | Gemma 4 26B A4B | |
|---|---|---|---|
| Best for | Fast reasoning and tool use at low cost | Coding, agents, structured output | Text and images, reasoning and tools at low cost |
| Context | ≈ 280 pages | ≈ 210 pages | ≈ 70 pages |
| Input / output, € per million | 0.12 / 0.30 | 0.10 / 0.30 | 0.12 / 0.30 |
| 1,000 chat answers | ~0.15 € | ~0.14 € | ~0.15 € |
Frequently asked questions
Is Nemotron 3.5 Lightning 30B free?
The model weights are open source; running them costs GPU time. On GPU Flow you pay €0.12 per million input tokens, and every new account gets €5 of credit when it adds a card.
What does the Nemotron 3.5 Lightning 30B API cost?
€0.12 per million input tokens and €0.30 per million output. VAT included, no monthly fee.
Does Nemotron 3.5 Lightning 30B support function calling?
Yes, with the standard OpenAI tools parameter.
Can I use Nemotron 3.5 Lightning 30B with Claude Code or Cursor?
Yes. Claude Code connects through the Anthropic-compatible endpoint (/v1/messages); Cursor and other tools through the OpenAI-compatible one (/v1/chat/completions).
Where does it run?
On our NVIDIA B200 GPUs, in a data centre in Spain.
Why is it called Lightning?
Because it is built for speed: a mixture of experts that activates only about 3 billion of its 30 billion parameters for each token.
Is it good for agents?
Yes. It reasons in several steps and supports tool calling, and its low price per call suits agents that run many times.