🔑 API Keys
Get free token → · Enable "Make calls to Inference Providers" in token permissions
Feeling cautious? Create a capped Project and generate your key inside it — an org-wide limit alone won't isolate this key from your others
Feeling cautious? Create a workspace with a spend cap and generate your key inside it — the Default Workspace can't be capped
Feeling cautious? Set a monthly spend cap in AI Studio’s Spend page before pasting it here
🧪 Select Models to Test
Pick models from the provider dropdowns below. Each option shows its published rate, and your selection is listed underneath — up to 5 at a time.
Nothing selected yet — choose up to 5 from the dropdowns above.
⚠️ Model Availability Notice: All paid API providers require at least a basic paid plan. OpenAI reasoning models (o3, o3-mini) use more tokens and may exhaust low-tier rate limits quickly. HuggingFace models are served via Inference Providers and free tier has rate limits.
Model access depends on your API key's plan tier. If a model returns an error, verify it is included in your provider plan.
Some models may require additional authorization or access approval from the provider before use. Check directly with OpenAI, Anthropic, Google, or NVIDIA to confirm your account has access to specific models.
A mix of accuracy and variety — good for most tasks