Suggestion: Adopt Sonnet 3.5 as Base Model—Or Provide Limited 3.5 Completions Alongside Unlimited GPT 4.1 (+ regular premium offer)

[techcow2]

Select Topic Area

Product Feedback

Copilot Feature Area

General

Body

I’d honestly consider coming back if you offered something like this. I hate to be “that guy,” and I know this might sound snobby, but I can’t help remembering one of my earliest interactions here: I brought up the rate limit issue, and a GitHub employee basically told me, “You could use another service instead of Copilot.” That response really stuck with me.

So, while I get the sense that customer retention isn’t a big priority, I wanted to put this out there—just in case someone cares. This kind of change could actually bring people like me back.

[Akash1134]

Hello @techcow2,

Thank you for sharing your feedback about GitHub Copilot's model offerings. Your suggestion to consider Claude Sonnet 3.5 as a base model or provide limited Sonnet completions alongside unlimited GPT-4.1 touches on important considerations about model choice, performance, and subscription value.

Current Model Landscape

I'm happy to share that GitHub has been expanding its AI model offerings significantly:

1. Claude Models: GitHub Copilot now supports multiple Claude models:

   • Claude 3.5 Sonnet: Available to all GitHub Copilot users, including those on the Free tier
   • Claude 3.7 Sonnet: Released in February 2025 and now available for Copilot Pro and Pro+ users
   • Claude Sonnet 4.0 and Claude Opus 4.0: Available for higher tier subscriptions

2. Model Choice: Copilot now offers a range of models that users can select based on their needs:

   • The Free tier provides access to Claude 3.5 Sonnet, GPT-4.1, and more, but with monthly limits (50 agent mode/chat requests, 2,000 completions)
   • The Pro tier ($10/month) offers unlimited agent mode and chat interactions with GPT-4.1, access to Claude 3.7/4.0 Sonnet and other models
   • The Pro+ tier ($39/month) provides access to all models, including premium ones like Claude 4.0 Opus and GPT-4.5

Your Feedback in Context

Your suggestion aligns with some of the changes GitHub has already implemented. Claude Sonnet models are now part of the model mix available to Copilot users, with the latest Claude 3.7 Sonnet offering improved performance over the 3.5 version. The rate limits you mentioned have been addressed in the tiered subscription model, with the Free tier providing a taste of the capabilities while the paid tiers remove these restrictions.

Resources for More Information

• GitHub Copilot Plans and Pricing - Details about the different subscription tiers and what each includes
• Using Claude in GitHub Copilot - Documentation on how to use Claude models in Copilot
• GitHub Changelog - Where you can track updates to GitHub services, including Copilot

Next Steps

I appreciate you taking the time to share your thoughts and previous experience. Customer feedback is incredibly valuable, and suggestions like yours help shape the evolution of GitHub Copilot. I'll make sure your feedback about model preferences and the importance of subscription flexibility is shared with the appropriate teams.

If you haven't explored Copilot recently, you might find that the current model selection and subscription options address some of your concerns. The Free tier now offers access to Claude 3.5 Sonnet (albeit with usage limits), while the Pro and Pro+ tiers provide access to the newer Claude 3.7 Sonnet and other advanced models.

Thank you again for your feedback and for being part of the GitHub community.