Models
357 sources · 1,263 models × 3,079 benchmarks · 162,127 scoresNot ranked · too few of its benchmarks are in the capability model to estimate a score
MAI Code 2 Flash is the smaller, faster MAI Code 2 model from Microsoft AI, for high-volume coding work such as explanations, targeted edits and routine fixes, released on November 17, 2026 in GitHub Copilot and as a limited-access preview in Microsoft Foundry through the OpenAI-compatible Responses API. Its model card describes a sparse mixture-of-experts transformer with an integrated vision encoder, with 12 billion active and about 398 billion total parameters, pre-trained from scratch on about 25 trillion tokens using logit distillation from a larger MAI teacher and sharing MAI Code 2's post-training. It accepts text and image input (922,000 input tokens) and returns text (128,000 output tokens) in a 1,050,000-token context, with low, medium, high, xhigh and max reasoning effort. Weights are not distributed, and Microsoft gives no per-token price, showing launch costs at a 50% promotional rate.
Scored on 4 benchmarks · 4 total scores · 0 independently reported
Organizations reporting headline scores for 4 benchmarks
| Software · 3 benchmarks | |||||
|---|---|---|---|---|---|
| CyberGymIn fitnot reported | 34 / 50 | ~4 mo | Jul 2026 | 2026-11-17 | |
| DeepSWE v1.1In fit | 44 / 81 | ~3 mo | ~Apr 2027 | 2026-11-17 | |
| Terminal-Bench 2.1In fit | 52 / 226 | ~4 mo | Nov 2026* | 2026-11-17 | |
| Science · 1 benchmark | |||||
| GPQA DiamondIn fitnot reported | 69 / 415 | ~5 mo | Apr 2026 | 2026-11-17 | |