The Glama Gateway API has no rate limits. We ensure all models remain available without usage restrictions. If you encounter any rate-limited model, please contact support@glama.ai.

Model aliases automatically point to the latest version of a model. For example, while the current GPT-4.5 version is "gpt-4.5-preview-2025-02-27", using the alias "gpt-4.5" ensures you'll always access the newest release when updates occur. Use aliases instead of specific version numbers to stay current.

When a model is available from multiple providers, you have two options: • Include the provider prefix: "fireworks/deepseek-chat-v3" • Use the model name only: "deepseek-chat-v3"

Using the model name alone enables automatic provider selection based on:

1. Success rate (reliability)
2. Cost (lowest pricing)
3. Response time (fastest performance)

For open-source models, we recommend omitting provider prefixes. This allows the gateway to dynamically route requests to the best-performing provider as availability and pricing fluctuate.

Using Glama Gateway provides several key advantages over direct provider access:

Better reliability and pricing: For models available from multiple providers, our intelligent routing ensures optimal performance and cost. The gateway automatically fails over to alternative providers during outages, maintaining service continuity.

No rate limits: We negotiate elevated rate limits with every provider and distribute this capacity across our user base, eliminating the throttling issues common with direct access.

Comprehensive observability: Track performance metrics and costs across all providers in one place. We offer detailed audit logs for every request, simplifying debugging, cost analysis, and compliance requirements.

Enhanced privacy: Your requests are anonymized within traffic from thousands of users. Unlike direct provider access where usage patterns can be tracked to your account, the gateway provides an additional layer of privacy protection.

Unified API: Access 100+ models from dozens of providers through a single, consistent interface—no need to manage multiple API keys, SDKs, or integration patterns.