name = "MAI-Code-1-Flash"
description = "Microsoft coding model built for fast, efficient assistance in everyday developer workflows"
family = "mai"
release_date = "2026-06-02"
last_updated = "2026-06-08"
attachment = false
reasoning = true
temperature = true
tool_call = true
structured_output = true
knowledge = "2025-12"
open_weights = false

[limit]
context = 256_000
output = 128_000

[modalities]
input = ["text"]
output = ["text"]

[[links]]
label = "Model card"
url = "https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF"
type = "model_card"

[[links]]
label = "Announcement"
url = "https://microsoft.ai/news/introducingmai-code-1-flash/"
type = "announcement"

[[benchmarks]]
name = "SWE-Bench Pro"
score = 51.2
metric = "resolve rate"
harness = "GitHub Copilot"
source = "https://microsoft.ai/news/introducingmai-code-1-flash/"
date = "2026-06-02"

[[benchmarks]]
name = "SWE-Bench Verified"
score = 71.6
metric = "resolved"
source = "https://llm-stats.com/benchmarks/swe-bench-verified"

[[benchmarks]]
name = "Terminal-Bench"
score = 54.8
metric = "success rate"
version = "2.0"
source = "https://llm-stats.com/benchmarks/terminal-bench-2"

[[benchmarks]]
name = "GPQA Diamond"
score = 84.6
metric = "accuracy"
source = "https://llm-stats.com/benchmarks/gpqa"