Compare GPT, Claude, Gemini, Llama, Qwen and more — using your own API keys at direct provider rates. The lab is free. Every query shows its cost before you run it.
Free • No signup • No credit card
NEW FEATURE🟢 NVIDIA Nemotron Lab — test Nemotron models with Nemotron Synthesis →💰 Free Lab
The lab itself is free. HuggingFace open-source models are also free — grab a token →
📐 Direct Rates
Your API calls go directly to OpenAI, Anthropic, and Google at their published per-token rates. You see the exact cost of every query.
🔒 Your Keys, Your Data
Keys live in your browser only — never stored server-side. Use Incognito for maximum security.
Everything you need to compare, fuse, and decide.
Test up to 5 models on the same prompt
Fuse multiple responses into one synthesized answer
Llama, Qwen, Gemma, GPT, Claude, Gemini & more
Bias, Toxicity, Safety, Hallucinations, PII & more
Upload documents as context for your prompts
Compare costs for all 22 models at a glance
Compare 5 models side-by-side, then fuse their answers into one. Direct API rates — see the total before you click fuse.
🔑 BYO Keys
Pay only for usage
🤗 Open Source
Mix proprietary and open models
💰 Pricing Guide
Full cost chart for all 22 models
Troubleshoot API key issues and check which models are live before running tests.
See Claude connect to live data sources in real-time using Anthropic's Model Context Protocol.
Paste a code segment and let multiple models evaluate it in parallel. Find bugs, get recommendations, and discover which model understands your code best.
Test NVIDIA's Nemotron models (MoE architecture) via NVIDIA's API. Free tier: 1,000 credits. Now with Nemotron Synthesis — run OpenAI, Anthropic & Google alongside Nemotron, then let Nemotron synthesize all results.