
NVIDIA has collaborated with OpenAI to optimize the new open-source gpt-oss-20b and gpt-oss-120b reasoning models for NVIDIA GPUs, enabling high-performance inference from the cloud to local RTX AI PCs. This initiative strengthens NVIDIA's AI leadership by democratizing advanced AI capabilities like chain-of-thought reasoning and long context lengths for developers, facilitating agentic AI applications. The optimized models, easily deployable via tools like Ollama, underscore NVIDIA's critical role in expanding AI accessibility and driving innovation in the AI ecosystem.
NVIDIA is reinforcing its full-stack dominance in the artificial intelligence sector through a strategic collaboration with OpenAI. The optimization of new open-source models, gpt-oss-20b and gpt-oss-120b, specifically for NVIDIA's GPU architecture extends its leadership from cloud-based training on H100s to on-device inference for its RTX AI PCs and workstations. This initiative is not merely a product launch but a strategic move to democratize advanced AI capabilities, such as long-context reasoning up to 131,072 tokens and efficient MXFP4 precision, for a broad developer community. By enabling high-performance local inference, with benchmarks reaching 256 tokens per second on the upcoming GeForce RTX 5090, NVIDIA is creating a significant demand driver for its high-end consumer hardware. The integration with popular developer tools like Ollama and Microsoft's AI Foundry Local further solidifies NVIDIA's ecosystem, ensuring its hardware remains the de facto standard for both professional and enthusiast AI development, thereby creating a powerful flywheel effect for its RTX product line.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request a DemoOverall Sentiment
strongly positive
Sentiment Score
0.85
Ticker Sentiment