MAI-Code-1.1-Flash

Claims and evidence

• Released 11 August 2026 ● Primary source[1]dateline [2]Model Summary: Release date [3]

• Status AvailableIn production in GitHub Copilot; no deprecation notice found up to 2026-10-01. ● Primary source[1]first paragraph [4] [3]Availability in GitHub Copilot

• Successor of MAI-Code-1-Flash ● Primary source[1]first paragraph [3]first paragraph [5] [6]README: New Release - MAI-Code-1.1-Flash
• Derived from (other) MAI-Thinking-1Training started from MAI-Thinking-1's compressed 5B-active-parameter mid-training checkpoint, followed by supervised fine-tuning, a further mid-training phase and reinforcement learning. ● Primary source[2]Training disclosure; Model dependencies

• Change · Modality Adds native image input; MAI-Code-1-Flash accepted text only.Compared with MAI-Code-1-Flash ● Primary source[3]first paragraph [2]Model Summary: Inputs
• Change · Efficiency Microsoft says it uses about 25% fewer tokens per task and streams tokens about 25% faster in GitHub Copilot.Compared with MAI-Code-1-Flash ● Primary source[1]
• Change · Other Microsoft puts the price at a quarter of MAI-Code-1-Flash; GitHub lists a 73% lower list price.Compared with MAI-Code-1-Flash ● Primary source[1] [3]
• Change · Other Microsoft focused this version on command-line tasks in GitHub Copilot CLI and on .NET work, based on developer feedback.Compared with MAI-Code-1-Flash ● Primary source[1]

• Input text, image ● Primary source[2]Model Summary: Inputs [3]first paragraph
• Output text ● Primary source[2]Model Summary: Outputs
• Context window 256K tokens tokensThe numeric value reads K as 1,000; the source does not say. ● Primary source[2]Model Summary: Context length
• Parameters 138B total, 5B activeSparse mixture-of-experts. ● Primary source[2]Model Summary: Parameters

Lineage

Variants

No variants recorded in this record.

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Microsoft on AI Radar →