One Billion Tokens in Four Days: Why AI Coding with Top Models Is Unsustainable

[FW2017]

🏷️ Discussion Type

Product Feedback

💬 Feature/Topic Area

Copilot in GitHub

Body

I pulled the plug on my GitHub Copilot Pro+ subscription on June 1, 2026 — without burning a single credit from my shiny new quota. Why? Because I could already smell the disaster brewing for anyone who codes seriously. Better to grab a refund than keep spinning the hamster wheel.

Switching to DeepSeek was like lifting the curtain: in just four days of plain Java and Rust work, I racked up one billion tokens. And no, I wasn’t running marathon agents — just quick searches and refactors that wrapped up in minutes. Even with DeepSeek’s generous caching (95% free!), the sheer absurdity of the token economy shines through.

Now imagine the same billion tokens on OpenAI or Anthropic, with little to no caching. That’s a $30,000 bill for GPT‑5.5, or $25,000 for Opus 4.8 — for four days of coding. At that point, AI isn’t competing with human programmers; it’s pricing them out.

Serious developers need predictable costs, not a slot machine where tokens fly out faster than you can blink. Yet I keep hearing that “enterprises are fine with it” and that GitHub Copilot is built for them. Sorry, but that’s nonsense. In my own company — a major financial institution in Asia — we’re already steering away from overpriced models. There’s even talk of pooling resources into shared datacenters running open‑source models like DeepSeek to slash costs further.

The irony? The so‑called future of coding is collapsing under the weight of its own token math.

[github-actions[bot]]

💬 Your Product Feedback Has Been Submitted 🎉

Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users.

Here's what you can expect moving forward ⏩

• Your input will be carefully reviewed and cataloged by members of our product teams. 
  • Due to the high volume of submissions, we may not always be able to provide individual responses.
  • Rest assured, your feedback will help chart our course for product improvements.
• Other users may engage with your post, sharing their own perspectives or experiences. 
• GitHub staff may reach out for further clarification or insight. 
  • We may 'Answer' your discussion if there is a current solution, workaround, or roadmap/changelog post related to the feedback.

Where to look to see what's shipping 👀

• Read the Changelog for real-time updates on the latest GitHub features, enhancements, and calls for feedback.
• Explore our Product Roadmap, which details upcoming major releases and initiatives.

What you can do in the meantime 💻

• Upvote and comment on other user feedback Discussions that resonate with you.
• Add more information at any point! Useful details include: use cases, relevant labels, desired outcomes, and any accompanying screenshots.

As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities.

Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐

[Usman-Amin-AI]

Thank you for sharing your perspective.

Your feedback highlights a concern that many developers are beginning to evaluate more closely: the relationship between AI-assisted development and cost predictability.

While usage-based pricing can make sense from a provider's perspective, the challenge is that software development workflows are often difficult to estimate in advance. Features such as agent-based coding, repository-wide analysis, refactoring, and iterative development can generate substantial model usage, making it difficult for individual developers and organizations to forecast costs accurately.

The comparison you raise between commercial frontier models and open-source alternatives is also becoming increasingly relevant. As open models continue to improve, organizations are naturally evaluating whether self-hosted or hybrid solutions can provide a more sustainable balance between capability, privacy, and cost.

That said, different users have different requirements. Some teams may prioritize access to the latest proprietary models and be willing to accept higher usage costs, while others may place greater emphasis on predictable budgeting and operational control. Ideally, platforms should accommodate both approaches.

I believe the broader concern raised here is not simply the cost of tokens themselves, but the lack of transparency and predictability surrounding consumption. Developers need clear visibility into how credits are used, how agent actions translate into costs, and what level of usage can reasonably be expected from everyday development tasks.

As AI becomes more deeply integrated into software engineering workflows, pricing models that are understandable, predictable, and aligned with real-world usage will likely play a significant role in long-term adoption.

Thank you for contributing your experience and perspective to the discussion.