AIToday
AI Watch (Impress)Published: Nov 16, 2026, 16:00 JST

35 launches lmaas.io for in-house AI on open models

Send on LINE
35 launches lmaas.io for in-house AI on open models

3 Key Points

  1. What happened

    35 began offering lmaas.io, which lets firms pick from over 1,000 open models such as Gemma, gpt-oss and Llama, add in-house data via RAG, and fine-tune with LoRA, all without sending data to external AI services.

  2. Why it matters

    Companies worried about putting internal documents or customer data into outside AI tools get a self-contained environment with data separation, encryption and licence management, priced from 64,000 yen per month (excluding tax), which may suit firms without dedicated AI engineers.

WHO IT HITSEnterprise IT teams and compliance staff at companies that want to use generative AI but cannot send internal documents or customer data to outside services are the main audience; firms without dedicated AI engineers may find the single-environment setup useful.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

The launch comes as companies widen their use of generative AI but run into two practical blocks: uncertainty over whether internal documents and customer data may be entered into outside AI services, and the difficulty of building and managing an AI environment with in-house IT staff alone. lmaas.io is positioned to close both gaps in one place, covering everything from model selection and internal data connection to additional training and permission management for users and external connections.

The platform leans on open models rather than a single vendor's API, offering more than 1,000 options such as Gemma, gpt-oss and Llama, with RAG included on all plans and LoRA-based additional training on dedicated GPUs. It also addresses licence concerns by letting customers check on-screen whether each model permits commercial use, additional training and redistribution, and by keeping a record of licence consent.

On infrastructure, 35 offers a shared-GPU plan that spreads high-performance GPUs across customers to lower initial cost, and a dedicated-GPU plan suited to stable response times and long training runs. Pricing runs from 64,000 yen to 560,000 yen per month (excluding tax) across five plans, with a 90-day minimum term and a 10% discount for annual payment.

FAQ
Which AI models can I use?
You can choose and switch among more than 1,000 open models, including Google's Gemma, OpenAI's gpt-oss and Meta's Llama; the models and sizes available differ by plan.
How much does it cost?
There are five plans priced from 64,000 yen to 560,000 yen per month (excluding tax), with initial fees from 96,000 yen to 750,000 yen (excluding tax), depending on model size and GPU type.
Is my data sent to outside AI services?
Training data and chat history are processed inside the lmaas.io environment and are not sent to external services such as ChatGPT, except when the customer configures external API or MCP connections.
AI Watch (Impress)Read Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Next articleGemini 4 Hydrogen, Carbon split Google's model lineup