The landscape of artificial intelligence model pricing is undergoing a substantial transformation, with major providers like GitHub Copilot, Anthropic, and OpenAI implementing significant changes, including shifts to usage-based billing and price hikes. This trend is leading some users to explore more cost-effective open-source alternatives.
GitHub Copilot is transitioning to a usage-based billing model, effective June 1, 2026. Under this new structure, all Copilot plans, including Pro and Pro+, will include monthly AI Credits, with usage calculated based on token consumption. "Agentic workflows have fundamentally changed Copilot’s compute demands," GitHub stated, noting that long-running, parallelized sessions now consume far more resources. The company also temporarily paused new sign-ups and tightened usage limits, while Opus models are no longer available in the standard Pro plans, being reserved for Pro+ subscribers.
Anthropic, developers of the Claude models, recently caused a stir among users regarding the availability of its powerful Opus models. While Claude Opus 4.7 is generally available across its Pro, Max, Team, and Enterprise plans, the company conducted a "small test on ~2% of new prosumer signups" where Claude Code, which leverages Opus, was temporarily removed from the $20-per-month Pro plan. This led to confusion, with Anthropic's head of growth, Amol Avasare, clarifying that the company is "looking at different options to keep delivering a great experience for users" due to fundamentally changed usage patterns. Pricing for Opus 4.7 starts at $5 per million input tokens and $25 per million output tokens for API usage.
OpenAI has also implemented a series of price increases for its advanced models. According to reports, the cost of their best model has risen from an initial $8 to $10, then to $12, $14, and now $30. These increases reflect the escalating costs of developing and running cutting-edge AI, particularly for advanced models like GPT-4 Turbo, which saw a 3x increase for input tokens and a 2x increase for output tokens in some instances.
These shifts have prompted a re-evaluation among developers and businesses. As one user, Lisan al Gaib, noted on social media, "> it was nice having experienced the golden times, but they are over now." The user highlighted the increasing costs and suggested that "> you can basically vibe-code everything now with open models like Kimi-K2.6, GLM-5.1 and DeepSeek-V4 for a fraction of the cost," pointing towards a growing interest in open-source alternatives as proprietary AI services become more expensive.
