New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
MAI-Thinking-1
About MAI-Thinking-1
MAI-Thinking-1 is an AI model developed by Microsoft, in the language category, released in 2026, made available as a proprietary (API-only) model. MAI-Thinking-1 is Microsoft AI's first in-house flagship reasoning model, a medium-sized Mixture-of-Experts with roughly.
On this page you'll find MAI-Thinking-1's full specifications, including a context window of 256K tokens, 9 benchmark results. Review provider pricing and benchmark scores below, or compare MAI-Thinking-1 head-to-head with other language models.
Links
Community
No reports yet. Own this model on some hardware? Be the first to confirm it runs.
No comments yet. Start the discussion.
MAI-Thinking-1 benchmark scores▶
| Benchmark | Category | Score | Variant |
|---|---|---|---|
| Terminal-Bench 2.0 | agentic | 46% | |
| LiveCodeBench | code | 87.7% | |
| SWE-bench Pro | code | 52.8% | |
| SWE-bench Verified | code | 73.5% | |
| MMLU-Pro | general | 85% | |
| MultiChallenge | general | 53% | |
| AIME 2025 | math | 97% | |
| GPQA Diamond | reasoning | 84.2% | |
| BFCL v3 | tool-use | 72% |
More models like MAI-Thinking-1
Frequently asked questions
What is MAI-Thinking-1? ▶
MAI-Thinking-1 is Microsoft AI's first in-house flagship reasoning model, a medium-sized Mixture-of-Experts with roughly It is tracked on GenAIList with its specifications, benchmark scores and provider pricing.
Who created MAI-Thinking-1? ▶
MAI-Thinking-1 was developed by Microsoft and released in 2026.
Is MAI-Thinking-1 open source or proprietary? ▶
MAI-Thinking-1 is a proprietary model. It is accessed through an API rather than by downloading the weights.
What is MAI-Thinking-1's context window? ▶
MAI-Thinking-1 supports a context window of 256K tokens.
How much does MAI-Thinking-1 cost? ▶
Pricing for MAI-Thinking-1 depends on the provider. See the providers table on this page for the latest API rates.
How does MAI-Thinking-1 perform on benchmarks? ▶
MAI-Thinking-1 is benchmarked across 9 evaluations on GenAIList, including Terminal-Bench 2.0 (46%). See the full benchmark table below and compare it with other models.
Know a great video about MAI-Thinking-1?
Add YouTube Video
Add Tweet/Discussion
Report an Issue
Help us maintain accurate information in our database.
Suggest an Edit
Tell us what needs to be changed or corrected for