madhavan02/llm-eval-harness

LLM Eval Harness — Model Comparison Dashboard

Interactive dashboard comparing llama-3.3-70b-versatile vs llama-3.1-8b-instant on Groq, evaluated on 10 AI/ML Q&A pairs.

Metrics shown

• —P50 / P95 latency and time-to-first-token (TTFT)
• —Tokens per second throughput
• —Exact match and semantic similarity
• —Hallucination rate and RAGAS faithfulness
• —Cost per query and total cost

Source

github.com/madhavanbalaji02/llm-eval-harness