Math multi-agent recipes cover actor/verifier and related multi-model math training workflows.
Run one example from the repo root:
bash examples/math-multi-agent/qwen3-8b-actor-verifier-m2po-delta-deepscaler-data/scripts/run_qwen3-8b-actor-verifier-m2po-delta-deepscaler-data.shComplete guidance: docs/en/recipes/multi-agent.md.
GPU Resources
These recipes default to one 8xH100 node.