Neither wins outright: Google: Lyria 3 Pro Preview is the pick when cost matters at scale, NVIDIA: Nemotron 3 Nano 30B A3B when you need tool / function calling.
| Spec | Google: Lyria 3 Pro Preview | NVIDIA: Nemotron 3 Nano 30B A3B |
|---|---|---|
| Provider | NVIDIA | |
| Input price (per 1M tokens) | $0 | $0.06 |
| Output price (per 1M tokens) | $0 | $0.24 |
| Blended price (3:1 input/output) | $0 | $0.10 |
| Cheaper than … of all models we track | 90% | 80% |
| Context window | 1.0M tokens (~1,573 pages) | 262K tokens (~393 pages) |
| Larger context than … of all models | 80% | 47% |
| Max output per response | 66K tokens (~49,200 words) | 236K tokens (~176,900 words) |
| Input types | text, image | text |
| Output types | text, audio | text |
| Free variant | Yes | Yes |
| Listed since | 2026-03-30 | 2025-12-14 |
Monthly cost at list price for five common workloads. Token counts per unit are stated so you can scale them to your own traffic. A workload that does not fit a model's context window is marked rather than priced.
| Workload (per month) | Google: Lyria 3 Pro Preview | NVIDIA: Nemotron 3 Nano 30B A3B | Difference |
|---|---|---|---|
| Customer-support chatbot 50,000 conversations — each a 6-turn conversation (1,500 in / 400 out tokens) | $0 | $9.30 | Google: Lyria 3 Pro Preview saves $9.30 |
| RAG search over documents 20,000 questions — each a question answered from ~8 retrieved passages (6,000 in / 500 out tokens) | $0 | $9.60 | Google: Lyria 3 Pro Preview saves $9.60 |
| Coding / tool-using agent 2,000 tasks — each a multi-step task with tool calls and file context (60,000 in / 6,000 out tokens) | $0 | $10.08 | Google: Lyria 3 Pro Preview saves $10.08 |
| Batch summarisation 10,000 documents — each a 6,000-word report summarised to a page (8,000 in / 600 out tokens) | $0 | $6.24 | Google: Lyria 3 Pro Preview saves $6.24 |
| Long-document analysis 500 documents — each a 150-page contract or codebase read in one pass (100,000 in / 2,000 out tokens) | $0 | $3.24 | Google: Lyria 3 Pro Preview saves $3.24 |
| Capability | Google: Lyria 3 Pro Preview | NVIDIA: Nemotron 3 Nano 30B A3B | What it lets you do |
|---|---|---|---|
| Tool / function calling | — | ✓ Yes | call your APIs and run agent loops |
| Structured output (JSON schema) | ✓ Yes | ✓ Yes | return JSON that validates against your schema |
| Reasoning / thinking mode | — | ✓ Yes | spend extra tokens thinking before answering hard problems |
| Image input | ✓ Yes | — | read screenshots, charts and scanned pages |
| Audio input | — | — | take speech or audio directly |
| File / PDF input | — | — | accept documents without your own parsing step |
| Image output | — | — | generate images, not just text |
| Built-in web search | — | — | answer from live web results |
| Seed (reproducible sampling) | ✓ Yes | ✓ Yes | repeat a generation for tests and evals |
| Log-probabilities | — | ✓ Yes | read token confidence for classification and scoring |
From each model's published parameters and input/output types. “—” means not advertised, not proven absent.
| Date | Model | Input / 1M | Output / 1M |
|---|---|---|---|
| 2026-10-08 | NVIDIA: Nemotron 3 Nano 30B A3B | $0.06 | $0.24 |
| 2026-09-22 | NVIDIA: Nemotron 3 Nano 30B A3B | $0.05 | $0.20 |
| 2026-09-17 | NVIDIA: Nemotron 3 Nano 30B A3B | $0.06 | $0.24 |
Neither wins outright: Google: Lyria 3 Pro Preview is the pick when cost matters at scale, NVIDIA: Nemotron 3 Nano 30B A3B when you need tool / function calling. In short: Google: Lyria 3 Pro Preview when cost matters at scale, you need to fit long inputs in one request, you need image input, you want the newer model; NVIDIA: Nemotron 3 Nano 30B A3B when you need tool / function calling, you need reasoning / thinking mode, you need log-probabilities, you generate long outputs in one go.
Google: Lyria 3 Pro Preview is cheaper: $0 vs $0.10 per 1M tokens at a typical 3:1 input/output mix, 100% less. For 50,000 support conversations a month that is $0 against $9.30.
Google: Lyria 3 Pro Preview costs $0 per 1M input tokens and $0 per 1M output tokens; NVIDIA: Nemotron 3 Nano 30B A3B costs $0.06 input and $0.24 output. Output tokens are billed separately, so long answers shift the gap.
Google: Lyria 3 Pro Preview: 1.0M tokens, about 1,573 pages, against 262K (about 393 pages) for NVIDIA: Nemotron 3 Nano 30B A3B.
Tool calling — Google: Lyria 3 Pro Preview: not advertised, NVIDIA: Nemotron 3 Nano 30B A3B: yes. JSON-schema structured output — Google: Lyria 3 Pro Preview: yes, NVIDIA: Nemotron 3 Nano 30B A3B: yes.
Only Google: Lyria 3 Pro Preview accepts image input; NVIDIA: Nemotron 3 Nano 30B A3B is text-only for input.
Google: Lyria 3 Pro Preview has a free variant (google/lyria-3-pro-preview); NVIDIA: Nemotron 3 Nano 30B A3B has a free variant (nvidia/nemotron-3-nano-30b-a3b:free). Free variants usually carry tighter rate limits.
Google: Lyria 3 Pro Preview can write up to 66K tokens (about 49,200 words) per response; NVIDIA: Nemotron 3 Nano 30B A3B up to 236K (about 176,900 words).
Google: Lyria 3 Pro Preview was listed on 2026-03-30; NVIDIA: Nemotron 3 Nano 30B A3B on 2025-12-14.
Google: Lyria 3 Pro Preview: price unchanged since we started tracking it on 2026-03-31. NVIDIA: Nemotron 3 Nano 30B A3B: input price raised 17% ($0.05 → $0.06/1M), output price raised 17% ($0.20 → $0.24/1M) on 2026-10-08 (3 price changes since 2026-03-10).
Full-length songs are priced at $0.08 per song. Lyria 3 is Google's family of music generation models, available through the Gemini API. With Lyria 3, you can generate high-quality, 48kHz...
Full Google: Lyria 3 Pro Preview specs →NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully...
Full NVIDIA: Nemotron 3 Nano 30B A3B specs →