# NVIDIA Touts Codex GPT-5.5 Gains _NVIDIA is integrating OpenAI's Codex, powered by GPT-5.5 and running on its own hardware, to accelerate complex engineering and research tasks._ **Published:** 2026-05-12 **Source:** https://www.startuphub.ai/ai-news/artificial-intelligence/2026/nvidia-touts-codex-gpt-5-5-gains --- NVIDIA is integrating OpenAI's Codex, now running on its own GB200 and GB300 infrastructure, to accelerate engineering and research. This advanced AI tool, powered by [GPT-5.5](https://openai.com/index/nvidia), is enabling faster development cycles and more complex AI experiments. NVIDIA HardwareCore From the article 4 mentionsNVIDIA is integrating OpenAI's Codex, now running on its own GB200 and GB300 infrastructure, to accelerate engineering and research.Autonomous SessionsContextFrom the article 3 mentionsThe system's ability to handle extended, autonomous sessions allows it to identify issues and suggest solutions beyond initial prompts.runs onOpenAI CodexCoreAI tool for complex engineering and researchFrom the article 9+ mentionsNVIDIA is integrating OpenAI's Codex, now running on its own GB200 and GB300 infrastructure, to accelerate engineering and research.GPT-5.5CoreAdvanced AI model powering Codex capabilitiesFrom the article 6 mentionsThis advanced AI tool, powered by GPT-5.5, is enabling faster development cycles and more complex AI experiments.Engineering WorkflowsEffectCodex as primary tool for coding challengesFrom the article 3 mentions"Codex is our go-to tool for complex engineering tasks, and with GPT-5.5, it surfaces bugs and gaps in my program that other models weren’t able to find," stated Dennis Hannusch, Senior Software Engineer.Production SystemsOutcomeInternal AI agents deploy tools across companyFrom the article 3 mentionsUsing Codex, Hannusch transformed an internal platform from an MVP to a production-ready system, enhancing scalability and reliability.Accelerated TasksEffectFaster development cycles and complex AI experimentsFrom the article 3 mentionsThis accelerated development was critical given internal privacy constraints.Bug DetectionEffectSurfaces bugs and gaps other models miss Engineers are using [Codex](/ai-news/artificial-intelligence/2026/openai-s-safety-playbook-for-codex) as a primary tool for intricate coding challenges and to run comprehensive machine learning experiments. The system's ability to handle extended, autonomous sessions allows it to identify issues and suggest solutions beyond initial prompts. "[Codex](/ai-news/artificial-intelligence/2026/openai-codex-gets-automation) is our go-to tool for complex engineering tasks, and with GPT-5.5, it surfaces bugs and gaps in my program that other models weren’t able to find," stated Dennis Hannusch, Senior Software Engineer. ## Building Production Systems NVIDIA's internal AI agents team is instrumental in deploying these tools across the company. Codex with GPT-5.5 has become their default for demanding engineering work. Hannusch highlighted its increased autonomy: "I’m able to go for long sessions with multiple compactions and find that it still performs with top accuracy and manages to keep the work in context." He also noted its skill in selecting appropriate tools and capabilities. Using Codex, Hannusch transformed an internal platform from an MVP to a production-ready system, enhancing scalability and reliability. The team also rapidly developed an internal podcast recording app, comparable to Riverside, in just hours. This accelerated development was critical given internal privacy constraints. The Codex desktop app's computer interaction capabilities even allowed for autonomous testing of video and audio recording features as they were built. "Codex has completely changed the threshold for what’s worth building," Hannusch added. ## Automating Research Workflows For NVIDIA's research divisions, Codex is streamlining the entire research loop. This includes identifying research areas, scripting experiments, and executing them on remote infrastructure. "GPT-5.5 has been a massive unlock as a creative partner, especially when it comes to knowledge work," said Shaunak Joshi, an AI researcher. He described using Codex as a research agent, feeding it large volumes of relevant papers. Joshi found GPT-5.5 to be particularly creative, helping trace evidence across research papers and visualizing concept connections through suggested knowledge graphs. This aided in hypothesis generation. Codex then generates the necessary scripts for training models on NVIDIA's machine learning infrastructure. The app's SSH support simplifies remote host login and setup, allowing researchers to manage extensive workloads from their laptops. Furthermore, Codex excels at code modernization. "If you have an old codebase that isn’t that performant, Codex is really good at machine translation. So a lot of folks are taking their Python repository, sending it to GPT-5.5, and it’s rewriting it into Rust and making it like 20X more efficient," Joshi explained. Codex is also proving invaluable for tasks such as code translation, with teams converting Python repositories to Rust for significant performance gains. "We’re just scratching the surface of what it can do," Hannusch concluded. "I’m really excited to keep building real systems and see how far it can go." --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory.