Fusion routes each task to the model that fits.Fusion routes each task to the model that fits, with a near-zero error rate. Every major model is a click away.
Pick a model for a chat, give each agent and routine its own, and set your workspace default. Any agent they bring in uses the same one.
Fusion by default, with Claude, OpenAI and Z.ai one click away. Max mode puts in extra effort, and if a provider goes down, the same model keeps running elsewhere.
Choose which models your workspace can use and set credit limits for each member. Fusion always stays on.
Fusion picks the right model for every task. Instead of running your agent's whole job on one monolithic model, it routes each step to the right class of model: fast workhorse models for routine execution, frontier models for deep reasoning. That mix is what lets agents run long and in parallel without paying frontier rates on every step.
No. On Auto, Fusion does the routing — you never manage the switch. If you want control, you can route tasks across model families or standardize on one; switching takes a setting, not a rebuild.
Standard costs $199/seat/month and includes 1× Fusion usage. Pro costs $499/seat/month and includes 5× Fusion usage. Both plans offer the same full feature set. The only difference is usage. Every workspace starts with a 7-day free trial.
You can purchase additional credits at $0.85 per credit, with volume discounts available. Your account menu shows how much of your monthly Fusion allowance remains, so you can track usage before reaching your limit.
Yes. The model picker offers Claude, GPT, GLM, and more. Non-Fusion models are billed separately at usage rates and aren't included in your monthly plan. You'll see a warning before switching.
Fusion usage is allocated per seat and pooled across your workspace. Your team shares one combined allowance rather than managing separate limits for each person. Inviting teammates is free during the trial. Once it ends, a workspace admin assigns paid seats.
Opus 5.5
GPT-6.1 Sol
GLM 5.3