DeepArt Labs Adds Public Tier for Restricted Nova-7 AI Model

DeepArt Labs on Tuesday released a public version of its Nova-7 AI model, a system the company once internally rated as “too effective for public access.” The release, announced through the company’s developer portal, adds a tier called Nova-7 Lite that includes hard-coded content filters and real-time usage auditing.

What Nova-7 Lite Can Do

The model handles multimodal inputs — text, images, and audio — and produces full software applications, research summaries, and creative works. Internal benchmarks showed Nova-7 outperforming earlier models on reasoning tasks by 42 percent. The public version limits autonomous actions: it cannot directly execute code or browse the web on its own.

Safety Features Added to the Public Release

DeepArt Labs embedded a consent-based monitoring layer. Every prompt runs through a classifier that blocks requests involving weapons, controlled substance synthesis, or social manipulation. A dashboard shows real-time grievance logs, and users can flag outputs for review. The company confirmed it will share anonymized abuse patterns with the Partnership on AI.

Why the Model Was Previously Restricted

Nova-7 first surfaced in closed-door demos in March 2025. Engineers at the time worried its ability to simulate human reasoning and generate persuasive content could be misused for large-scale disinformation. The lab’s internal risk assessment, leaked to the press last year, rated the system “unacceptable for general release without external oversight.”

Industry Reaction and Security Debates

The launch arrives amid wider scrutiny of unconstrained AI tools. Facial recognition on smart glasses has already sparked privacy debates, and the arrival of Nova-7 Lite adds to pressure on lawmakers to update AI safety rules. Separately, Microsoft’s AI chief recently declared the company’s independence from OpenAI to chase superintelligence, signaling that industry efforts to build advanced models are accelerating.

DeepArt Labs’ Next Steps

The company plans to keep the full Nova-7 model under enterprise-only access. A developer preview of an API with higher rate limits is scheduled for August 2026. Paid tiers will add priority support and custom fine-tuning options. DeepArt said it will submit the Lite version for third-party audit by September.

Frequently Asked Questions

How to get started with DeepArt Labs Nova-7 public tier?

Visit the DeepArt Labs website and create a free account. The Nova-7 public tier is automatically available; simply select the Nova-7 model from the art generation menu and begin creating with its restricted feature set.

What is the restricted Nova-7 AI model by DeepArt Labs?

The Nova-7 AI model is DeepArt Labs' advanced image generator, now offered in a limited public tier. This restricted version allows users to test core capabilities like style transfer and basic generation with capped resolution and daily usage limits.

Why can I only use Nova-7 with limitations on DeepArt Labs?

The Nova-7 model is computationally intensive, so DeepArt Labs applies usage caps in the public tier to manage server load and ensure fair access. Full unrestricted access is reserved for paid plans, but the public tier still lets you explore the model's core strengths.

Is the DeepArt Labs Nova-7 public tier completely free to use?

Yes, the Nova-7 public tier is free with no credit card required. You get a set number of generations per day and standard-definition outputs; upgrading to a premium plan unlocks higher limits and full resolution.

How does Nova-7 public tier compare to DeepArt Labs premium models?

The public tier offers a taste of Nova-7's style versatility but with lower resolution and daily caps, while premium tiers include the unrestricted Nova-7 model plus additional models like UltraRender. Advanced users will notice faster processing, batch generation, and commercial licensing only in paid plans.