fix: Information Integration evaluation - handle multiple answer variants with pipe-separated format

feat: Show all 9 LLM models in app dropdown, add comprehensive code review and metric analysis documentation

feat: Add Moonshot Kimi K2 Instruct model to DEFAULT_MODELS

Update all LLMs to new model list with 4 default models

Replace decommissioned llama-3.1-70b-versatile with llama2-70b-4096

Reduce RPM limit to 25 (safe margin below 30) and increase request interval to 2.5s

Implement RPM rate limiting (30 requests/minute) with sliding window

Add deepseek-r1-distill-llama-70b free chat model

Replace decommissioned mixtral-8x7b-32768 with llama-3.1-70b-versatile

Add 4th model: meta-llama/llama-4-maverick-17b-128e-instruct

Replace deprecated gemma2-9b-it with mixtral-8x7b-32768

Deploy RGB Metrics dashboard - 2026-01-04 22:05:37