Nemotron 3 Nano 30B A3B
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B
Mixture-of-experts model, strong on reasoning traces.
Coding models
Nine coding models served from one gateway. Point any OpenAI-compatible client at a single base URL, pick a model per request, and change providers by editing one line.
Catalog
9 models. 1M-token context on the flagship tier.
Served through Fireworks AI and Nebius AI Studio infrastructure.
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B
Mixture-of-experts model, strong on reasoning traces.
google/gemma-3-27b-it
Efficient open-weight model, strong on general instructions.
glm-5.3-flash
Fast tier-1 model with strong instruction following.
deepseek-v4p1-flash
Fastest and cheapest tier-1 workhorse. Good first opinion.
gpt-oss-120b
Open-weight 120B reasoning model, careful with math and code.
minmax-m3
Balanced frontier model, strong summarisation and synthesis.
glm-5.3
Frontier GLM tier. Strong long-form writing and structure.
deepseek-v4-pro
Frontier DeepSeek tier. Deepest analysis, slowest and priciest.
kimi-k3
Long-context specialist, good at very large pasted context.
Each plan includes models up to a usage tier. Requesting a model outside your plan returns HTTP 403 with the list of models you can use.
| Model | FreeFree | Starter$2/mo | Go$5/mo | Pro$15/mo |
|---|---|---|---|---|
| Nemotron 3 Nano 30B A3Bnvidia/NVIDIA-Nemotron-3-Nano-30B-A3B | — | Included | Included | Included |
| Gemma 3 27Bgoogle/gemma-3-27b-it | — | Included | Included | — |
| GLM 5.3 Flashglm-5.3-flash | — | Included | Included | Included |
| DeepSeek V4.1 Flashdeepseek-v4p1-flash | Included | Included | Included | Included |
| GPT-OSS 120Bgpt-oss-120b | — | — | Included | Included |
| MiniMax M3minmax-m3 | — | — | — | Included |
| GLM 5.3glm-5.3 | — | — | — | Included |
| DeepSeek V4 Prodeepseek-v4-pro | — | — | — | Included |
| Kimi K3kimi-k3 | — | — | — | Included |