implement AgentClinic benchmark

environments/agentclinic/README.md

@@ -0,0 +1,61 @@
+# AgentClinic - Responses API
+
+Multi-agent medical diagnosis environment using OpenAI's Responses API for reasoning models.
+
+## Installation
+
+```bash
+# Install the environment
+uv pip install -e environments/agentclinic/
+
+# Set your API key
+export OPENAI_API_KEY="your-key-here"
+```
+
+## Quick Start
+
+### MedQA Dataset (214 cases)
+
+```bash
+uv run --active -m verifiers.scripts.eval \
+  -m o4-mini \
+  -b https://api.openai.com/v1 \
+  -k OPENAI_API_KEY \
+  agentclinic_response \
+  -n 10 \
+  --rollouts-per-example 2 \
+  --max-concurrent 2 \
+  -T 0.0 \
+  -s \
+  --env-args '{"dataset_path": "agentclinic_medqa_extended.jsonl"}'
+```
+
+### NEJM Dataset (119 cases)
+
+```bash
+uv run --active -m verifiers.scripts.eval \
+  -m o4-mini \
+  -b https://api.openai.com/v1 \
+  -k OPENAI_API_KEY \
+  agentclinic_response \
+  -n 10 \
+  --rollouts-per-example 2 \
+  --max-concurrent 2 \
+  -T 0.0 \
+  -s \
+  --env-args '{"dataset_path": "agentclinic_nejm_extended.jsonl"}'
+```
+
+
+### Environment Args
+
+```bash
+--env-args '{
+  "dataset_path": "agentclinic_medqa_extended.jsonl",
+  "max_turns": 20,
+  "patient_model": "gpt-4o-mini",
+  "measurement_model": "gpt-4o-mini",
+  "moderator_model": "gpt-4o-mini"
+}'
+```
+