From Cloud Wars · Tom Smith · · 1m 20s
GitHub's HydraFusion picks AI models for you at runtime
New feature automatically orchestrates multiple AI models to match coding tasks, balancing performance and cost without manual selection.
In brief
New feature automatically orchestrates multiple AI models to match coding tasks, balancing performance and cost without manual selection. GitHub's Project HydraFusion automatically selects and orchestrates the best AI models for each coding task at runtime, eliminating the need for developers to manually choose…
AI picks the right model for you
GitHub's Project HydraFusion orchestrates AI models at runtime, matching developers' tasks to the best model automatically.
No more manual model coordination
Developers typically choose models manually, asking one to review another's work or escalating tough problems to more capable models.
From selection to full orchestration
GitHub's previous Auto Model Selection let developers pick "Auto" for tasks. HydraFusion goes further, creating full execution plans at runtime.
Three execution patterns
Single uses one model for speed. Cascade drafts with one model, then escalates if needed. Critique has one model draft while another reviews.
Smart cost management
HydraFusion only calls extra models when it expects better results, choosing the simplest workflow to meet each developer's needs.
“HydraFusion only calls in extra models when it's likely to improve the result”
Available now in Copilot CLI
Project HydraFusion is available on all GitHub Copilot plans through /experimental in the CLI. Usage is based on tokens consumed.
The takeaway
GitHub's Project HydraFusion automatically selects and orchestrates the best AI models for each coding task at runtime, eliminating the need for developers to manually choose models.
