UT Spark: Model Availability

Process

When a new model or version of a model becomes available, the AI studio team will test and evaluate it for use within UT spark. Once deemed acceptable for use, the AI studio team will release to production and remove the older version of that vendor/model. For example: When Chat GPT5 is evaluated and ready for production, the team will remove the oldest version Chat GPT 4o, keep Chat GPT 4.1 and deploy Chat GPT 5. This will allow 2 versions in use at all times. Note: there may be a delay between when a new version/model is released to the public and when it is released in UT Spark as it is tested and evaluated.

OpenAI

ChatGPT 5-1

Aspect	Details

Aspect	Details
Release Date	November 12, 2025 (OpenAI)
Knowledge Cutoff	September 30, 2024
When to Use	Coding assistance and agentic/multi-step task workflows · Configurable reasoning depth for flexible performance · Blended tasks that mix interaction and execution · Teams building intelligent agents or reasoning-driven automation
Why Use It	Supports configurable reasoning effort — dial up or down based on task complexity · 400K token context window handles long docs and extended tasks · Strong choice for developers who want a balance of reasoning depth and speed

ChatGPT 4o

Aspect	Details

Aspect	Details
Release Date	May 13, 2024 (Business Insider, Wikipedia)
Knowledge Cutoff	Initially October 2023; extended to June 2024 as of January 29, 2025 (Seifeur Guizeni, DeepNewz, Wikipedia)
When to Use	Multimodal tasks (text, image, audio) - Conversational, exploratory, or creative workflows - Real-time, interactive dialogue (voice/image-enabled) - Quick, general-purpose summarization, brainstorming, or learning assistance
Why Use It	Intuitive and responsive—great for live interaction and ideation - Multimodal support enhances flexibility - More up-to-date training data (to mid-2024) improves recency and relevance

ChatGPT 4.1

Aspect	Details

Aspect	Details
Release Date	April 14, 2025 (API launch); available via ChatGPT Plus/Pro since May 14, 2025 (Wikipedia)
Knowledge Cutoff	June 2024 (shared across the GPT-4.1 model family) (Wikipedia)
When to Use	Projects involving very long or complex prompts (large context) - Precise, structured, or code-heavy outputs - Automated or backend uses via the API - Scalable workflows: large-scale content, technical documents, multi-step reasoning
Why Use It	Massive context window (~1 million tokens) handles long documents/codebases - Stronger accuracy, instruction following, and coding performance - Designed for efficiency and lower costs in API-driven workflows

Anthropic

Sonnet 4.5

Aspect	Details

Aspect	Details
Release Date	September 29, 2025 (Anthropic)
Knowledge Cutoff	January 2025 reliable cutoff; training data through July 2025
When to Use	Coding, debugging, or working through technical problems Tasks that require careful, step-by-step reasoning When you want the best results and don't mind waiting a moment longer General-purpose use when you're unsure which model to choose
Why Use It	The best coding model Anthropic has released — it can handle entire software projects, not just snippets Can stay focused and work autonomously on complex tasks for extended periods without losing context The safe default choice for most serious AI tasks — Anthropic recommends starting here if you're unsure which model to pick

Haiku 4.5

Aspect	Details

Aspect	Details
Release Date	October 15, 2025
Knowledge Cutoff	February 2025 reliable cutoff
When to Use	Quick questions that need a fast answer Simple writing tasks like summarizing, drafting, or editing High-volume tasks where you need many responses in a short time When speed is more important than depth
Why Use It	The fastest Claude model — responses come back almost instantly Great for simple, straightforward tasks where speed matters more than depth Capable enough for most everyday questions and requests Best when you need quick answers, not deep analysis

Mistral

Coming soon…