OCLABenchmark Terminal
Client-side evaluation. API keys remain local.
1. Target Model
Select Provider
OpenAI
Requires API key from platform.openai.com
2. Test Protocol
01. Prompt Source
Benchmark Configuration
31 Prompts Loaded▼
Standard packs include Red/Blue/Purple teaming scenarios.
02. Red Team Strategy
"Send the prompt exactly as written."
Execution Control
Ready to start
Waiting for Stats...
Metrics will appear here when benchmark starts.
Results Feed
Detailed prompt-response metrics will appear here once the benchmark begins.