DeepInfra raises $107M Series B to scale the inference cloud — read the announcement

DeepInfra Launches Access to NVIDIA Nemotron Models for Vision, Retrieval, and AI Safety

Published on 2025.10.28 by Yessen Kanapin
DeepInfra Launches Access to NVIDIA Nemotron Models for Vision, Retrieval, and AI Safety

DeepInfra is serving the new, open NVIDIA Nemotron vision language and OCR AI models from day zero of their release. As a leading inference provider committed to performance and cost-efficiency, we're making these cutting-edge models available at the industry's best prices, empowering developers to build specialized AI agents without compromising on budget or performance.

The NVIDIA Nemotron model Family

NVIDIA Nemotron represents a paradigm shift in enterprise AI development. This comprehensive family of open models, datasets, and technologies unlocks unprecedented opportunities for developers to create highly efficient and accurate specialized agentic AI. What sets Nemotron apart is its commitment to transparency—offering open weights, open data, and tools that provide enterprises with complete data control and deployment flexibility.

Nemotron Models on DeepInfra platform

Nemotron Nano 2 VL - 12B Multimodal Reasoning Powerhouse

This 12-billion parameter model leverages a hybrid Mamba-Transformer architecture to deliver exceptional accuracy in image and video understanding and document intelligence tasks. With industry-leading performance on OCRBench v2 and an average score of 73.2 across multiple benchmarks, Nemotron Nano 2 VL represents a significant leap forward in multimodal AI capabilities.

Nemotron Parse 1.1 - Efficient Information Extraction

The 1-billion parameter vision-language model specializes in accurate parsing of complex documents including PDFs, business contracts, financial statements, and technical diagrams. Its efficiency makes it ideal for high-volume document processing workflows.

Complete Nemotron Ecosystem

DeepInfra is providing access to the entire Nemotron family, including NVIDIA Nemotron Safety Guard for culturally-aware content moderation and the Nemotron RAG collection for intelligent search and knowledge retrieval applications.

Why DeepInfra is Your Ideal Nemotron Partner

Performance-Optimized Infrastructure

We run on our own cutting-edge NVIDIA Blackwell inference-optimized infrastructure in secure data centers. This ensures you get the best possible performance and reliability for your Nemotron deployments. Define your latency and throughput targets and we'll architect a solution to meet your needs.

Cost-Effective Scaling

Our low pay-as-you-go pricing model means you can scale to trillions of tokens without breaking the bank. No long-term contracts, no hidden fees—just simple, transparent pricing that grows with your needs.

Developer-First Approach

We've designed our APIs for maximum developer productivity with hands-on technical support to ensure your success. Whether you're optimizing for cost, latency, throughput, or scale, we design solutions around your specific priorities.

Enterprise-Grade Security and Privacy

With our zero-retention policy, your inputs, outputs, and user data remain completely private. DeepInfra is SOC 2 and ISO 27001 certified, following industry best practices in information security and privacy.

Getting Started with NVIDIA Nemotron on DeepInfra

Visit our Nemotron page to explore our competitive rates for Nemotron inference, or check out DeepInfra docs to learn more about our complete model ecosystem and developer resources. The future of specialized AI agents is here, and it's more accessible than ever through the powerful combination of NVIDIA Nemotron open models and DeepInfra's inference platform. Join us in building the next generation of intelligent applications.

Related articles
Step 3.5 Flash API Benchmarks: Latency, Throughput & CostStep 3.5 Flash API Benchmarks: Latency, Throughput & Cost<p>About Step 3.5 Flash Step 3.5 Flash is an open-weights reasoning model released in February 2026 by StepFun. It leverages a sparse Mixture of Experts (MoE) architecture with 196 billion total parameters and only 11 billion active parameters per token during inference — delivering state-of-the-art performance at a fraction of the cost of dense models. [&hellip;]</p>
Kimi K3 vs Claude Opus 4.8 vs GPT-5.6 Sol: Practical AI Model ComparisonKimi K3 vs Claude Opus 4.8 vs GPT-5.6 Sol: Practical AI Model Comparison<p>What&#8217;s changed since this comparison first ran Two of the three models compared here have since been succeeded, and GPT-5.6 Sol&#8217;s pricing has changed materially. Three of the strongest models available on DeepInfra right now don’t separate cleanly by capability tier. Kimi K3 (available via DeepInfra), Claude Opus 4.8, and GPT-5.6 Sol all score within [&hellip;]</p>
From Screenshots to Tool Calls: NVIDIA Nemotron 3.5 Super VL on DeepInfraFrom Screenshots to Tool Calls: NVIDIA Nemotron 3.5 Super VL on DeepInfraNVIDIA Nemotron 3.5 Super VL, the new open multimodal model in the Nemotron family, is available on DeepInfra from day zero. It brings visual understanding, reasoning, and tool calling together in one model, through an OpenAI-compatible API with pay-per-token pricing.