From Startup Fortune · · 1m
Microsoft AI Model Beats OpenAI with Less Compute
Microsoft AI chief claims model outperforms OpenAI using 25% less compute.
In brief
Microsoft AI chief claims model outperforms OpenAI using 25% less compute. Microsoft's MAI-Base-2 model beats OpenAI's GPT-5.6 with 25% less compute, using 12x less than DeepSeek. Originally reported by Startup Fortune.
Microsoft AI Model Outperforms OpenAI
Mustafa Suleyman says Microsoft's MAI-Base-2 model beats OpenAI's GPT-5.6 system.
25% Less Compute Used
MAI-Base-2 used about a quarter of the compute that OpenAI spent training its comparable system.
Direct Benchmarking
Microsoft privately benchmarked MAI-Base-2 against OpenAI model on pattern understanding tests.
12x Less Than DeepSeek
MAI-Base-2 used 12 times less computation to reach comparable result with DeepSeek.
Microsoft's Financial Stake
Microsoft holds $135 billion stake in OpenAI Group PBC.
AGI Partnership Shift
Microsoft and OpenAI ended AGI clause partnership after 2026 agreement.
Economic Impact
Microsoft's inference margins on MAI-1 run near 100% gross vs 60-70% on OpenAI API.
The takeaway
Microsoft's MAI-Base-2 model beats OpenAI's GPT-5.6 with 25% less compute, using 12x less than DeepSeek.
