New Prompt Playground, Evals, And Integration Support
- Prompt Playground: Now supports GPT-4.5 & Anthropic Sonnet 3.7 and Thinking Budgets
- Instrumentation: SmolagentsInstrumentor to trace smolagents by Hugging Face
- Evals: o3 support, Audio & Multi-Modal Evaluations
- Integrations: Phoenix now supports LiteLLM Proxy & Cleanlabs evals
Improvements and Bug Fixes 🐛
- Admin: Show percent used of DB
- Configuration: Add environment variable for allocated DB storage capacity
- Tracing: Delete selected traces
- Tracing: Make trace tree more readable on smaller sizes
- Experiments: Ensure type is correct on
run_experiment - Experiments: Allow experiment run JSON downloads
- Python Client: Add anthropic thinking config param
- Components: Add ToggleButton