Understand this story
Pydantic AI is now available as a supported agent in Scale AI's open-source AgentEnv Framework. According to the official Pydantic AI article, the agent can run multi-step tasks in reproducible environments such as email, Slack, and help desks, and is scored by validators. It runs in-process and uses the Coder capabilities of the Pydantic AI Harness, including shell, read_file, write_file, edit_file, list_files, grep, and delegate_task, supports any OpenAI-compatible endpoint, and can use the vLLM provider by prefixing the model name with vllm:. Environment tools are exposed via MCP and agents connect via A2A; for the same task, you only need to replace a2a_agent_id with claude-code to compare different agents.
AI-generated from public reports
Latest development · Report
Pydantic AI Becomes an Officially Supported Agent in the AgentEnv FrameworkDevelopments
1 confirmed developments
Pydantic AI becomes an agent supported by AgentEnv Framework
Source:English text is being prepared.
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 6
Official Pydantic AI articleFirst-party
Pydantic AI Becomes an Officially Supported Agent in the AgentEnv Framework
Pydantic AI is now integrated into Scale AI's open-source AgentEnv Framework as a supported agent, capable of running multi-step tasks in reproducible environments such as email, Slack, and help desks, with scoring by validators. The agent runs in-process and uses the Coder capabilities of the Pydantic AI Harness (shell, read_file, write_file, edit_file, list_files, grep, delegate_task), supports any OpenAI-compatible endpoint, and can use the vLLM provider by prefixing the model name with vllm:. Environment tools are exposed via MCP and agents connect via A2A; for the same task, you only need to change a2a_agent_id to claude-code to compare different agents.