Right this moment, we’re thrilled to announce that OpenAI’s GPT‑5.4 is now usually accessible in Microsoft Foundry: a mannequin designed to assist organizations transfer from planning work to reliably finishing it in manufacturing environments. As AI brokers are utilized to longer, extra advanced workflows; consistency and observe‑by means of turn into as vital as uncooked intelligence. GPT‑5.4 combines stronger reasoning with in-built pc use capabilities to help automation eventualities, and reliable execution throughout instruments, recordsdata, and multi‑step workflows at scale.
GPT-5.4 is designed for organizations working AI in actual manufacturing environments, the place consistency, instruction adherence, and sustained context are essential to success. The mannequin brings collectively advances in reasoning, coding, and agentic workflows to assist AI methods not solely plan duties however full them with fewer interruptions and decreased handbook oversight.
In contrast with earlier generations, GPT-5.4 emphasizes stability throughout longer interactions, enabling groups to deploy agentic AI with larger confidence in day-to-day manufacturing use.
GPT-5.4 introduces developments that intention for manufacturing grade AI:
- Extra constant reasoning over time, serving to keep intent throughout multi‑flip and multi‑step interactions
- Enhanced instruction alignment to scale back immediate tuning and oversight
- Latency improved efficiency for responsive, real-time workflows
- Built-in pc use capabilities for structured orchestration of instruments, file entry, knowledge extraction, guarded code execution, and agent handoffs
- Extra reliable device invocation lowering immediate tuning and human oversight
- Greater‑high quality generated artifacts, together with paperwork, spreadsheets, and shows with extra constant construction
Collectively, these enhancements help AI methods that behave extra predictably as duties develop in size and complexity.
GPT‑5.4 delivers sensible worth throughout a variety of manufacturing eventualities the place observe‑by means of and reliability are important:
- Agent‑pushed workflows, resembling buyer help, analysis help, and enterprise course of automation
- Enterprise data work, together with doc drafting, knowledge evaluation, and presentation‑prepared outputs
- Developer workflows, spanning code era, refactoring, debugging help, and UI scaffolding
- Prolonged reasoning duties, the place logical consistency have to be preserved throughout longer interactions
Groups profit from decreased activity drift, fewer mid‑workflow failures, and extra predictable outcomes when deploying GPT‑5.4 in manufacturing.
GPT‑5.4 Professional, a premium variant designed for eventualities the place analytical depth and completeness are prioritized over latency.
Further capabilities embody:
- Multi‑path reasoning analysis, permitting various approaches to be explored earlier than deciding on a ultimate response
- Higher analytical depth, supporting issues with commerce‑offs or a number of legitimate options
- Improved stability throughout lengthy reasoning chains, particularly in sustained analytical duties
- Enhanced resolution help, the place rigor and thoroughness outweigh pace issues
Organizations usually choose GPT‑5.4 Professional when deeper evaluation is required resembling scientific analysis and sophisticated issues, whereas GPT‑5.4 stays the proper alternative for workloads that prioritize dependable execution and agentic observe‑by means of.
GPT‑5.4 and GPT‑5.4 Professional can be found by means of Microsoft Foundry, which offers the operational controls organizations have to deploy AI responsibly in manufacturing environments. Foundry helps coverage enforcement, monitoring, model administration, and auditability, serving to groups handle AI methods all through their lifecycle.
By deploying GPT‑5.4 by means of Microsoft Foundry, organizations can combine superior agentic capabilities into present environments whereas aligning with safety, compliance, and operational necessities from day one.
Buyer Highlight
GPT‑5.4 units a brand new bar for manufacturing‑prepared AI by combining stronger reasoning with reliable execution. By enterprise‑grade deployment in Microsoft Foundry, organizations can transfer past experimentation and confidently construct AI methods that full advanced work at scale. Pc use capabilities will probably be launched shortly after launch.
GPT‑5.4 <272K enter tokens context size in Microsoft Foundry is priced at $2.50 per million enter tokens, $0.25 per million cached enter tokens, and $15.00 per million output tokens. The GPT‑5.4 >272K enter tokens context size in Microsoft Foundry is priced at $5.00 per million enter tokens, $0.50 per million cached enter tokens, and $22.50 per million output tokens. The GPT-5.4 is on the market at launch in Normal International and Normal Information Zone (US), with further deployment choices coming quickly. GPT‑5.4 Professional is priced at $30.00 per million enter tokens, and $180.00 per million output tokens, and is on the market at launch in Normal International.
Construct brokers for real-world workloads. Begin constructing with GPT‑5.4 in Microsoft Foundry in the present day.
