AI Center: Support for new models

The AI Center now supports the latest flagship models from Anthropic, OpenAI, and Google Gemini, organizes every model into four clear price tiers, and retires deprecated model IDs. Cost-optimizing existing automations and upgrading to flagship reasoning for complex work both get a lot easier.
‍

Key highlights

• Latest Flagship Models: Claude Opus 4.8, GPT 5.5, and Gemini 3.1 Pro (2M-token context) are available for the most demanding reasoning, extraction, and analysis tasks across the full Anthropic, OpenAI, and Gemini catalogs.
• Four-Tier Pricing Structure: Every model is classified as Budget (up to two dollars per MTok output), Standard (two to six dollars), Premium (six to twenty dollars), or Flagship (over twenty dollars), so picking the right model for a workflow no longer requires researching each provider's pricing page separately.
• Retired Models Removed: Anthropic's Sonnet 4, Opus 4, Sonnet 3.7, Sonnet 3.5, Claude 3 Haiku, and Claude 3 Opus are no longer callable on the API and have been removed from the catalog. Gemini 2.0 Flash and Flash-Lite are also removed.
• Tier Revisions on Kept Models: Several models have moved between tiers to match current provider pricing (Sonnet 4.5 up from Standard to Premium; GPT 5.1 and GPT 5 down from Flagship to Premium; GPT 5 Mini and GPT 4o Mini down to Budget). Existing configurations continue to work; the tier badge is what changed.
  ‍

How it works

• Open the AI Center in Workspace Administration and browse the refreshed catalog; models are grouped by provider (Anthropic, OpenAI, Gemini) and labeled with tier badges.
• Pick the right model for a workflow based on tier: Budget for tagging and lightweight extraction, Standard for everyday summaries and structured output, Premium for content generation and analysis, Flagship for the hardest multi-step reasoning.
• Configure the model on any AI Field Agent, AI Assist prompt, or automation-driven AI action; existing configurations continue to work unless they reference a removed model ID (in which case update to the recommended replacement).
• Use the tier badges to spot cost-optimization opportunities: high-volume automation runs using a Premium model for a task that a Standard or Budget model could handle just as well are the easiest wins.
  ‍

Applies to every AI-powered surface in SmartSuite: AI Field Agents, AI Assist, and automation-triggered AI actions.