Switchyard
Open weightNVIDIA · Released September 2026
Switchyard is an open-source model router that switches between multiple models to optimize the cost of requests.
It belongs to NVIDIA's open-weight Nemotron line, tuned for helpfulness and agents on top of leading open architectures.
Facts and pricing
- Provider
- NVIDIA
- Context window
- 1M tokens
- Input price
- -1.2e+7 credits / 1M tokens ($-1.0e+6 raw)
- Output price
- -1.2e+7 credits / 1M tokens ($-1.0e+6 raw)
- Vision (image input)
- No
- Tool calling
- No
- Extended reasoning
- No
Credits are what Zeplik bills: 1 credit = $0.10, computed from the raw provider rate with a 1.20x margin. Raw prices shown per 1M tokens.
What Switchyard is best for
- Everyday chat and drafting on an open-weight model with transparent lineage
- Very long documents and codebases: a 1M-token window fits entire books or repositories in one conversation
Example prompts
Prompts that suit a open weight model like Switchyard:
- Draft a project update from these rough notes
- Explain how HTTPS works to a curious teenager
- Turn this list of features into a changelog entry
- Brainstorm objections to this proposal and how to answer them
NVIDIA family
| Model | Context | Input cr/M | Output cr/M | Released |
|---|---|---|---|---|
| Nemotron 3.5 Super VL | 1.0M | 1.8 | 5.6 | October 2026 |
| Switchyard | 1M | -1.2e+7 | -1.2e+7 | September 2026 |
| Nemotron 3.5 Lightning | 262K | 0.84 | 2.4 | August 2026 |
| Nemotron 3.5 Content Safety | 131K | 2.4 | 2.4 | June 2026 |
| Nemotron 3 Ultra | 262K | 6.0 | 26.4 | June 2026 |
| Nemotron 3 Super | 262K | 0.90 | 4.8 | March 2026 |
| Nemotron 3 Nano 30B A3B | 262K | 0.72 | 2.9 | December 2025 |
Model comparisons
- Claude Fable 5.2 vs Claude Fable 5.1
- Claude Opus 5.5 vs Claude Fable 5.2
- GPT-6 Astra vs Claude Fable 5.2
- Claude Opus 5.5 vs Claude Opus 5
Frequently asked questions
- What is Switchyard?
- Switchyard is an AI model by NVIDIA. Switchyard is an open-source model router that switches between multiple models to optimize the cost of requests.
- How much does Switchyard cost on Zeplik?
- Input tokens cost -1.2e+7 credits per million and output tokens -1.2e+7 credits per million (1 credit = $0.10; the raw provider rates are $-1.0e+6 and $-1.0e+6 per million). New accounts start with free credits.
- How long can a conversation with Switchyard be?
- Switchyard has a 1M-token context window (1,000,000 tokens), which covers the conversation plus any documents you attach.
- Does Switchyard support images and tools?
- Switchyard is text-only and does not support tool calling.
Related models
NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total.
NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total.
NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B.
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE).
NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.
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.
More you can do with Switchyard
- 468 AI skillsReady-to-run expert methods for writing, data, code and more, each running on Switchyard or any model you pick.
- 1582 integrationsConnect Gmail, Slack, GitHub, Notion and hundreds more so the assistant can work in your apps.
- Simple pricingPlus, Pro and Max plans with a monthly credit allowance that works across every model.