Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context. It’s
No signup required — 10,000 session credits for guests. 50,000 free credits on account creation.
Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context. It’s post-trained for agentic workflows (RAG, tool calling) via SFT across math, code
131,072-token context window — enough for long documents.
Extremely low per-token price compared to frontier models — good for high-volume workloads.
Hosted by NVIDIA — you can access it here alongside GPT-4o, Claude, Gemini and 100+ more on one plan.
Switch to any other model mid-conversation from the picker.
Around 270–675 credits per typical message. 15M Pro credits buy roughly 22,222–55,555 messages.
NVIDIA: Llama 3.3 Nemotron Super 49B V1.5 is a chat/completion model served by NVIDIA and accessed through Faceb.ai. Try it without signing up — guest users get 10,000 session credits to experiment.
You get 50,000 credits on signup — usually enough for a handful of messages. After that, Pro is $14.99/month for 15M credits, or top-ups from $5.
131,072 tokens. Paste your source material in, no need to truncate.
Yes. Any API key from /account/api/ works with model slug `nvidia/llama-3.3-nemotron-super-49b-v1.5`. The OpenAI SDK works with base_url=https://api.faceb.ai/v1.
Depends on the task. Faceb.ai lets you switch models per message — benchmark side-by-side by asking both the same prompt, which is more reliable than abstract comparisons.
Around 270–675 credits per typical message. 15M Pro credits buy roughly 22,222–55,555 messages.
No. We contractually request that upstream providers not train on content routed through us. Your chat history lives only on your account.
It depends on the model size and training mix. For serious code work the developer favourites are Claude 3.5 Sonnet and DeepSeek V3; for quick edits, most capable models work fine.
Yes — point your SDK at https://api.faceb.ai/v1 and use this model's slug as the model parameter. Everything else works as normal.
Check the model catalog — multimodal models are marked in the picker. If it accepts images, you can drop screenshots and diagrams straight into the chat.
Yes — the picker is always at the top of the chat. Previous context carries over.
Yes. Our catalog auto-fetches from the upstream aggregator, so provider updates and new versions appear in the picker as soon as they're available.
Your Faceb.ai credits work for every model — switch per message, no extra subscriptions.
OpenAI's flagship multimodal model — text, vision, and code.
Chat with GPT-4o →OpenAI's fastest, cheapest frontier model — great default.
Chat with GPT-4o mini →Anthropic's best balance of quality and cost — a coder favourite.
Chat with Claude 3.5 Sonnet →One subscription covers every frontier model — switch between them per message. No extra API keys, no extra bills.
Start chatting with NVIDIA: Llama 3.3 Nemotron Super 49B V1.5 → Go Pro · $14.99/mo