Adds image input and a 256K-token context, solves more terminal tasks with fewer tokens than MAI-Code-1-Flash, and Microsoft says it costs a quarter as much to run — available to every GitHub Copilot user.
Compared with MAI-Code-1-Flash · the model it replaces
Finishes 6 in 10 multi-step tasks at the command line
Terminal Bench 2.1 — running real commands to complete a task, inside GitHub Copilot
62.9%
▲ up from 51.7% · +11 points
Slightly better at fixing real software bugs, using a fifth fewer tokens
SWE-bench Verified — fixing real bugs in real software projects
72.6%
▲ up from 71.6% · +1 point · 8.6K tokens per task, was 10.8K
Costs a quarter of what the previous version did
Price — Microsoft's stated cost vs MAI-Code-1-Flash; exact API rate not yet published
1/4 the cost
▲ was $0.75 / $4.50 per 1M tokens
MAI-Code-1.1-Flash is the current model in this line. The card shows how long it has been since the last release against the line's usual gap and its recent ones — the rhythm, not a date.
Newest first. See the full line with every card →
One email a week: what AI models shipped, the actual differences vs the previous version, and what's coming next. No filler.