Does GitHub have any plans for improvement on rate limits while using anthropic models.

[Raj-Koneru]

Select Topic Area

Question

Body

Our teams have encountered rate limits when using anthropic models during our regular development activities, which can occur when providers are experiencing high loads. Although the rate limits typically resolve after a few minutes, they can sometimes significantly disrupt our workflows. Hence interested to understand if GitHub has any plans for improvement on rate limits.

[meeran-jeelani]

Rate limits for third-party models (including Anthropic) are primarily enforced by the model provider and can vary based on current load, usage tier, and capacity. GitHub typically applies provider-recommended limits to ensure platform stability. While temporary spikes can cause short-term throttling, improvements usually depend on upstream provider capacity and GitHub’s internal optimizations. For now, implementing client-side retries, backoff strategies, and usage smoothing is the recommended approach.

[beenish-1]

Rate limits for third-party models like Anthropic happen mainly when providers experience high demand. GitHub applies the recommended limits to keep the service stable for everyone.

While these limits usually reset quickly, we know they can interrupt workflows. GitHub monitors usage and looks for ways to improve responsiveness, but many improvements depend on the provider’s capacity.

In the meantime, it helps to:

• Add client-side retries with backoff
• Spread requests over time
• Use batching or asynchronous calls when possible

Your feedback is valuable and helps guide future improvements.