Wednesday, November 18, 2026

Microsoft (MS): “Claude-Level Coding AI at One-Tenth the Price”... Unveils 7 New AI Models [Global AI Briefing]

Input
2026-11-18 08:34:20
Updated
2026-11-18 08:34:20
[Financial News] Microsoft (MS) has made a full-fledged entry into the AI model market, putting price competition front and center.
In the high-performance AI market led by OpenAI and Anthropic, Microsoft has introduced models whose coding performance is close to that of competing models but whose prices are up to 95% lower.
Microsoft AI (MAI) announced that it was accelerating efforts to build an independent AI ecosystem, unveiling seven AI models at once for business needs including coding, cybersecurity, image generation, voice, and medical record creation at its annual developer event, Ignite 2026, held at Chase Center in San Francisco, United States, on the 17th (local time).
MAI emphasized that it had lowered the price of its latest AI models relative to their performance, helping reduce costs even when AI agents run 24 hours a day.
As more companies use AI, inference costs incurred when models are deployed in real-world operations are emerging as an important competitive factor, on a par with peak model performance. The strategy is to lead the market through price competitiveness.
Microsoft AI (MAI) announced that it was accelerating efforts to build an independent AI ecosystem, unveiling seven AI models at once for business needs including coding, cybersecurity, image generation, voice, and medical record creation at its annual developer event, Ignite 2026. Satya Nadella, chief executive officer (CEO) of Microsoft (MS), is presenting. (Source: Yonhap News Agency)

MAI said its software development AI, MAI Code 2, delivers 98% of the performance of OpenAI’s GPT-6.1 Sol at about 40% of the price, based on the average across four coding benchmarks. Compared with Anthropic’s Claude Opus 5.5, it delivers 96% of the performance at just 5% of the price.
That works out to prices about 60% and 95% lower, respectively, than those of the competing models.
MAI also unveiled a model specializing in cybersecurity. MAI Cyber 2 Flash is designed to have AI identify software vulnerabilities and support security personnel in responding to them. MAI said its own evaluations showed that it could deliver quality on par with OpenAI’s GPT-6 Astra at less than 2% of the cost.
MAI says its image editing model, MAI Image 2.7, ranked first in relevant evaluations. Its real-time voice conversation model, MAI Realtime 1, is available starting at $0.01 per minute of audio. MAI also says its medical voice documentation model is being used to create records of 10 million doctor-patient consultations a month.
MAI also unveiled MAI Licensed Weights, which allows companies to customize AI models using their own data and run them directly in the Microsoft Azure cloud environment. The aim is to move beyond simply renting AI models and enable companies to control their own business data and how they operate AI.
With this announcement, MS signaled its intention to launch full-scale price competition in the global AI market.
Mustafa Suleyman, head of MAI, emphasized, “What I keep hearing from chief information officers (CIOs) and chief financial officers (CFOs) is that spending on tokens is soaring,” and added, “At current prices for frontier models, it is difficult to sustain the cost structure.” Suleyman went on to say, “We won’t pursue specialized capabilities at the expense of putting a cost burden on customers,” and pointed out, “If we build the same models to push the frontiers of science, people end up paying for those capabilities even when fixing a single everyday bug.”
[email protected] Lee Gu-sun Reporter