20/3 Weekly Update: New Models, Free Nemotron, and Organizations

Weekly Product Update: New Models, Free Nemotron, and Team Accounts

It’s been an exciting week in AI. At NVIDIA GTC, Jensen Huang said “the inflection point of inference has arrived” - and we’re seeing that play out in real time in how teams are starting to use Doubleword, with more workloads moving away from real-time constraints and toward large-scale async systems.

This week, we’ve shipped a few updates to make that even easier and significantly cheaper.

New model support: Qwen3.5-4B & Nemotron-3-Super (120B)

We’ve added two new models at opposite ends of the spectrum, giving you more flexibility depending on your workload. Qwen3.5-4B is a compact reasoning model with a 262K context window, comparable to GPT-OSS-20B but at a fraction of the cost, making it a strong fit for high-volume batch workloads and agent pipelines where efficiency matters.

Pricing
High Priority: $0.05 / $0.08
Standard Priority: $0.04 / $0.06

Nemotron-3-Super (120B) is a high-performance agentic model built for coding, planning, and long-context reasoning, well suited to more complex workloads like eval pipelines, document processing, synthetic data generation, and multi-step agents.

To coincide with GTC, we’ve extended the promotion - it’s free on the Standard tier through the weekend. If you’ve had a workload you’ve been meaning to run at scale, this is the moment to do it.

Pricing
High Priority: $0.23 / $0.56
Standard Priority: Free

New: Organizations (team accounts)

We’ve also rolled out organizations, making it easier to run workloads as a team. You can now share credits across users, manage usage centrally, and get full visibility into workloads and spend, making it much easier to coordinate usage under a single account.

To get access, you can request it directly in the console via the support button (“Create organization request”), or reach out to us and we’ll fast-track it.

Try it at scale

Between the new models and the free Nemotron window, it’s a great time to properly test workloads at scale, whether that’s evals, document processing, synthetic data, or agent pipelines. We’d love to see what you run.