1 month ago
28
Today, we’re thrilled to denote that OpenAI’s GPT‑5.4 is present mostly disposable successful Microsoft Foundry: a exemplary designed to assistance organizations determination from readying enactment to reliably completing it successful accumulation environments. As AI agents are applied to longer, much analyzable workflows; consistency and follow‑through go arsenic important arsenic earthy intelligence. GPT‑5.4 combines stronger reasoning with built successful machine usage capabilities to enactment automation scenarios, and dependable execution crossed tools, files, and multi‑step workflows astatine scale.
GPT-5.4 is designed for organizations operating AI successful existent accumulation environments, wherever consistency, acquisition adherence, and sustained discourse are captious to success. The exemplary brings unneurotic advances successful reasoning, coding, and agentic workflows to assistance AI systems not lone program tasks but implicit them with less interruptions and reduced manual oversight.
Compared with earlier generations, GPT-5.4 emphasizes stableness crossed longer interactions, enabling teams to deploy agentic AI with greater assurance successful day-to-day accumulation use.
GPT-5.4 introduces advancements that purpose for accumulation people AI:
- More accordant reasoning implicit time, helping support intent crossed multi‑turn and multi‑step interactions
- Enhanced acquisition alignment to trim punctual tuning and oversight
- Latency improved performance for responsive, real-time workflows
- Integrated machine usage capabilities for structured orchestration of tools, record access, information extraction, guarded codification execution, and cause handoffs
- More dependable instrumentality invocation reducing punctual tuning and quality oversight
- Higher‑quality generated artifacts, including documents, spreadsheets, and presentations with much accordant structure
Together, these improvements enactment AI systems that behave much predictably arsenic tasks turn successful magnitude and complexity.
GPT‑5.4 delivers applicable worth crossed a wide scope of accumulation scenarios wherever follow‑through and reliability are essential:
- Agent‑driven workflows, specified arsenic lawsuit support, probe assistance, and concern process automation
- Enterprise cognition work, including papers drafting, information analysis, and presentation‑ready outputs
- Developer workflows, spanning codification generation, refactoring, debugging support, and UI scaffolding
- Extended reasoning tasks, wherever logical consistency indispensable beryllium preserved crossed longer interactions
Teams payment from reduced task drift, less mid‑workflow failures, and much predictable outcomes erstwhile deploying GPT‑5.4 successful production.
GPT‑5.4 Pro, a premium variant designed for scenarios wherever analytical extent and completeness are prioritized implicit latency.
Additional capabilities include:
- Multi‑path reasoning evaluation, allowing alternate approaches to beryllium explored earlier selecting a last response
- Greater analytical depth, supporting problems with trade‑offs oregon aggregate valid solutions
- Improved stableness crossed agelong reasoning chains, particularly successful sustained analytical tasks
- Enhanced determination support, wherever rigor and thoroughness outweigh velocity considerations
Organizations typically prime GPT‑5.4 Pro erstwhile deeper investigation is required specified arsenic technological probe and analyzable problems, portion GPT‑5.4 remains the close prime for workloads that prioritize reliable execution and agentic follow‑through.
GPT‑5.4 and GPT‑5.4 Pro are disposable done Microsoft Foundry, which provides the operational controls organizations request to deploy AI responsibly successful accumulation environments. Foundry supports argumentation enforcement, monitoring, mentation management, and auditability, helping teams negociate AI systems passim their lifecycle.
By deploying GPT‑5.4 done Microsoft Foundry, organizations tin integrate precocious agentic capabilities into existing environments portion aligning with security, compliance, and operational requirements from time one.
Customer Spotlight
GPT‑5.4 sets a caller barroom for production‑ready AI by combining stronger reasoning with dependable execution. Through enterprise‑grade deployment successful Microsoft Foundry, organizations tin determination beyond experimentation and confidently physique AI systems that implicit analyzable enactment astatine scale. Computer usage capabilities volition beryllium introduced soon aft launch.
GPT‑5.4 <272K input tokens discourse magnitude successful Microsoft Foundry is priced astatine $2.50 per cardinal input tokens, $0.25 per cardinal cached input tokens, and $15.00 per cardinal output tokens. The GPT‑5.4 >272K input tokens discourse magnitude successful Microsoft Foundry is priced astatine $5.00 per cardinal input tokens, $0.50 per cardinal cached input tokens, and $22.50 per cardinal output tokens. The GPT-5.4 is disposable astatine motorboat successful Standard Global and Standard Data Zone (US), with further deployment options coming soon. GPT‑5.4 Pro is priced astatine $30.00 per cardinal input tokens, and $180.00 per cardinal output tokens, and is disposable astatine motorboat successful Standard Global.
Build agents for real-world workloads. Start gathering with GPT‑5.4 in Microsoft Foundry today.