Model Selection

Agentforce Vibes lets you choose which AI model handles your requests. The models you see depend on which AI provider your organization has configured.

Click the model name in the bottom-left corner of the Chat interface (for example, “Opus 4.6”) to open the model selector. Available models appear grouped by provider. Select a model to start using it immediately.

The model selection persists across your session until you change it again. You can switch models mid-conversation without losing context.

Models generally fall into three tiers. For example, with the Claude Code provider:

• Most capable (Opus 4.6)—best for complex multi-file tasks, architectural planning, and debugging subtle issues. Reasons more deeply but responds more slowly and consumes more tokens.
• Balanced (Sonnet 4.6)—good for everyday development tasks where you need quality output at reasonable speed.
• Fastest (Haiku 4.5)—best for quick questions, simple code generation, and explanation requests. Responds quickly but may produce less thorough output on complex tasks.

Other providers offer similar tiers with their own models. Choose the tier that matches your current task. You don’t need the most capable model for every interaction.

• The models available to you depend on your AI provider configuration and license type.
• Different models consume tokens at different rates. More capable models typically use more tokens per response.
• Switching models doesn’t reset your conversation or context.