Requesty
Fireworks AI

nemotron-lightning-3.5-30b-a3b

Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.

ReasoningTool callingCachingJSON schema
CompareDocs

Specifications

Context window262K tokens
Max output262K tokens
API typechat
AddedAug 15, 2026
Model ID
Data retentionNo
Used for trainingNo
Provider location🇺🇸 US
Privacy policyFireworks Privacy Policy→

Benchmarks

Benchmarks haven't been published yet for this exact variant.

Some variants (region-specific deployments, highspeed tiers) share benchmarks with their base model. Check the base model page or the Fireworks AI models overview.

Pricing

Prices updated August 18, 2026
Input / 1M
$0.05
Output / 1M
$0.20
Cache write
N/A
Cache read / 1M
$0.01
Estimated cost
100K input + 10K output$0.0070
1M input + 100K output$0.0700
10M input + 1M output$0.70

Requesty charges exactly what the upstream provider charges, no markup, no per-request fees. Prompt caching and smart routing can reduce effective cost by 30-80%.

Quickstart

Drop-in compatible with the OpenAI SDK. Change the base URL, swap in your Requesty API key, and set the model to fireworks/nemotron-lightning-3.5-30b-a3b.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_REQUESTY_API_KEY",
    base_url="https://router.requesty.ai/v1",
)

response = client.chat.completions.create(
    model="fireworks/nemotron-lightning-3.5-30b-a3b",
    messages=[
        {"role": "user", "content": "Explain quantum computing in one paragraph."},
    ],
)

print(response.choices[0].message.content)

Other Fireworks AI models

deepseek-v4-pro-0813
1M ctx$1.32/1M
Coding 69
muse-glimmer-30b
131K ctx$0.35/1M
qwen3.8-max
1M ctx$2.00/1M
Coding 72
deepseek-v4-flash-0731
1M ctx$0.14/1M
Coding 69
kimi-k3
1M ctx$3.00/1M
Coding 76
glm-5.2-fast
1M ctx$2.10/1M

Frequently asked questions

How much does nemotron-lightning-3.5-30b-a3b cost?
nemotron-lightning-3.5-30b-a3b is priced at $0.05 per million input tokens and $0.20 per million output tokens when accessed via Requesty. Prompt caching is supported, which can cut effective input cost by up to 90% on repeated context. Requesty charges exactly what the upstream provider charges, we don't add markup.
What is the context window of nemotron-lightning-3.5-30b-a3b?
nemotron-lightning-3.5-30b-a3b has a context window of 262K tokens, with a maximum output of 262K tokens per response. That's roughly 350 words of input you can fit in a single prompt.
What can nemotron-lightning-3.5-30b-a3b do?
nemotron-lightning-3.5-30b-a3b supports tool calling, extended reasoning, prompt caching, structured outputs (JSON schema). You can call it through any OpenAI-compatible client by pointing base_url to Requesty.
How do I use nemotron-lightning-3.5-30b-a3b with the OpenAI SDK?
Install the OpenAI SDK, set base_url to "https://router.requesty.ai/v1", set your API key to your Requesty key, and set the model to "fireworks/nemotron-lightning-3.5-30b-a3b". The Quickstart above shows Python, JavaScript and cURL snippets.
Can I run nemotron-lightning-3.5-30b-a3b through Requesty?
Yes. nemotron-lightning-3.5-30b-a3b runs through Requesty's OpenAI-compatible API, served from Fireworks AI. You do not host the model yourself: point base_url at Requesty, set the model to "fireworks/nemotron-lightning-3.5-30b-a3b", and requests are routed to the upstream provider with automatic failover. The same key gives you 600+ other models too.

Access nemotron-lightning-3.5-30b-a3b through Requesty

One API key, 600+ models, OpenAI-compatible. No markup on provider prices, automatic failover, and smart caching built-in.

All Fireworks AI models