Tony Kim
Feb 19, 2026 20:58
GitHub deprecates three AI fashions from Copilot on Feb 17, pushing customers to Claude Opus 4.6 and GPT-5.2. Enterprise admins should replace mannequin insurance policies.
GitHub has pulled the plug on three AI fashions powering its Copilot coding assistant, forcing builders emigrate to newer alternate options because the platform continues its aggressive mannequin refresh cycle.
The deprecations, efficient February 17, 2026, have an effect on Claude Opus 4.1, GPT-5, and GPT-5-Codex throughout all Copilot experiences together with Chat, inline edits, agent modes, and code completions. GitHub is pointing customers towards Claude Opus 4.6, GPT-5.2, and GPT-5.2-Codex as replacements.
A part of a Broader AI Housecleaning
This transfer follows OpenAI’s personal retirement of a number of ChatGPT fashions simply 4 days earlier on February 13. The corporate pulled GPT-5, GPT-4o, GPT-4.1, and OpenAI o4-mini from the ChatGPT interface, citing remarkably low adoption—solely about 0.1% of every day customers had been nonetheless touching GPT-4o.
GPT-4o’s journey has been notably rocky. OpenAI briefly withdrew it through the GPT-5 rollout in August 2025, then introduced it again after consumer complaints, solely to kill it once more this month. Criticism over “sycophantic” conduct—the mannequin’s tendency to inform customers what they needed to listen to—apparently did not assist its case.
Anthropic has been working a parallel cleanup. Claude 3.5 Sonnet and its v2 variant hit their shutdown date on February 19, whereas Claude 3.5 Haiku faces a July 5 cutoff after being deprecated in early January.
What Admins Must Do
Enterprise prospects cannot simply sit this one out. Copilot Enterprise directors have to allow entry to the alternative fashions by way of their mannequin insurance policies in Copilot settings. As soon as configured, the brand new fashions will seem within the Chat mannequin selector in VS Code and on github.com.
The deprecated fashions will disappear robotically—no guide removing required. But when your workforce’s workflows or integrations reference particular mannequin variations, these will break with out updates.
GitHub Enterprise prospects with issues ought to contact their account managers. For everybody else, the documentation on supported fashions has been up to date to mirror the modifications.
The speedy deprecation cycle highlights an uncomfortable actuality of constructing on high of AI infrastructure: the muse retains shifting. Builders who hardcoded mannequin preferences into their workflows are studying that lesson once more.
Picture supply: Shutterstock
