Model details
GPT Chat Latest
GPT Chat Latest is presented on the Microsoft Foundry catalog as a preview, multimodal conversation model aimed at enterprise applications that need natural, context-aware dialogue. The listing describes it as powering fast chat experiences with adaptive reasoning, tuned instruction-following, and built-in safety guardrails, with the ability to apply step-by-step reasoning when a query benefits from a deeper answer rather than a short reply. Because it is delivered through the Direct from Azure portfolio, the model is managed by Microsoft, billed alongside other Foundry resources, and accessible through standard Azure deployment and governance workflows, including pay-as-you-go scaling and reserved PTU options for predictable capacity.
In practical terms, the model is positioned as a drop-in chat backend for customer support bots, internal assistants, and other conversational surfaces where responsiveness and safe, on-policy responses matter more than open experimentation. Its adaptive reasoning lets teams rely on concise answers for routine turns while still getting more thorough explanations on harder questions, and the multimodal positioning makes it suitable for workflows that combine text with attached images or documents. The Foundry catalog page carries a version stamp of 2026-06-24, indicating the listing is actively maintained, which is useful signal for teams planning pilots or proofs of concept on Azure before committing to a general-availability tier.
Quick Info
Powered by- Provider
- Azure
- Model key
- gpt-chat-latest
- Release date
- May 5, 2026
- Last updated
- May 28, 2026
- Knowledge cutoff
- 2025-12-01
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $5.00
- Output token cost
- $30.00
Limits
- Input tokens
- 111,616 tokens
- Output tokens
- 16,384 tokens
- Context window
- 128,000 tokens