M-1.5 Launches: AskAIs Moves from Generating Content to Completing Work with Super Agents

Published: June 2025

Today we are launching the next AskAIs model system: M-1.5, powered by the M1.5 intelligence core.

Since the first M-1 launched in 2023, we have kept asking one question: Can AI do more than answer questions, write content, and generate images—can it actually help people complete work?

The first M-1 already provided chat, text generation, image generation, and APIs. But as AskAIs served more founders, independent developers, and small teams, we learned that users rarely need one more answer. They need a complete result.

Launching a product means research, positioning, business models, development plans, website content, marketing assets, contracts, receipts, expenses, publishing schedules, progress tracking, and investor or partner materials. Those tasks are connected, span multiple tools, and depend on persistent context.

M-1.5 is therefore not simply designed to produce better answers. It is designed to understand an objective, break down complex work, use the right capabilities, process files and information, and keep moving under the user's control.

M-1.5 is a super-agent model system for founders, independent developers, and modern teams.

What are M-1.5 and M1.5?

M1.5 is the intelligence core responsible for language understanding, task reasoning, context, planning, and result generation.

M-1.5 is the complete agent system built on that core. It includes multi-step planning, agent execution, file and document handling, tool orchestration, project context, result checking, team workflows, startup work modules, and user authorization controls.

M1.5 thinks and understands; M-1.5 turns that intelligence into practical work.

From chat model to work agent

Traditional chat follows “question → answer → end.” M-1.5 follows “objective → context → task breakdown → plan → materials → tools → results → gap check → continued progress.”

If a user asks for early planning for an AI file-management product, M-1.5 can organize positioning and users, analyze problems, define core features, create development phases, draft the site and FAQ, prepare documentation and launch checks, plan initial marketing, identify contract and finance materials, and organize everything into reusable project assets.

The user can inspect, change, or stop the work at any time without manually coordinating every intermediate step.

Why focus on startup work?

Early-stage founders often handle product, marketing, sales, finance, documents, support, and team management at once. Receipts need classification, contracts need summaries, meetings need follow-up tasks, websites need updates, social channels need content, and investor materials need revision.

Each task may take only a short time, but together they repeatedly interrupt the work that requires judgment. M-1.5 does not replace a founder's decisions. It takes on research, organization, generation, checking, and coordination so founders can protect their attention.

Core super-agent capabilities

Understand the complete objective

“Prepare our product launch” is not one writing task. It may include product information, site copy, launch checks, marketing, notifications, help content, risk review, and post-launch measurement. M-1.5 builds a task structure around the desired outcome.

Break down multi-step work

It turns large objectives into executable steps, recognizes dependencies, and identifies where information or approval is required. A marketing plan, for example, should follow positioning, audience, and budget discovery.

Maintain project context

Product material, brand voice, market information, team feedback, and previous decisions can continue to inform later work instead of being explained from scratch in every conversation.

Process documents and files

M-1.5 can help organize business plans, product requirements, contracts, company files, invoices, receipts, meeting notes, customer feedback, research, drafts, and project lists. It can summarize, classify, extract key details, and use the material to create further deliverables.

Call tools to complete work

Within the user's authorization, the agent can research public information, analyze tables, create documents, generate visuals, organize tasks, write code, inspect web content, and work with project files. The model chooses the method; tools carry out the operation.

Check and revise results

The first output is not automatically the finish line. M-1.5 checks for missing sections, structural issues, contradictions, inconsistent data or formats, usability, approval points, and obvious risks, then continues improving the result.

What startup work can M-1.5 support?

Business and market research: competitor comparisons, opportunities, user pain points, positioning, differentiation, and risk analysis.

Product planning: objectives, user scenarios, core features, version scope, priorities, roadmaps, launch checks, and feedback processes.

Content and brand: homepages, feature pages, blog posts, help content, social plans, email, announcements, advertising ideas, and brand guidelines.

Documents and administration: receipts, invoices, contracts, meetings, and company files. Legal, tax, and accounting judgments still belong with qualified professionals.

Finance preparation: receipt classification, expense organization, monthly packets, missing-document checks, reimbursement preparation, and accountant handovers—not accounting decisions.

Fundraising and partnerships: company profiles, product overviews, markets, business models, roadmaps, FAQ, meeting materials, follow-up, and due-diligence checklists.

Engineering: requirements, technical plans, code changes, error analysis, architecture checks, API docs, test plans, pre-deployment checks, and handover material.

How does one task get completed?

For “We are launching in one month; help us prepare,” M-1.5 first learns the product, users, current progress, existing site and content, team roles, budget, and objectives. It can then organize product completion, testing, site copy, help center, notifications, social content, launch assets, analytics, support readiness, rollback planning, and post-launch review.

A conventional chat model tells the user what to do. M-1.5 attempts to continue doing that work under the user's control.

What changed from M-1 to M-1.5?

• from individual chats to complex project context;
• from one-off text and images to multi-step workflows;
• from limited files to document and project processing;
• from answers to task planning and tool orchestration;
• from repeated prompting to objective-driven continuation;
• from a first result to gap checking and revision;
• from a generative AI platform to a startup work super agent.

The user remains in control

Greater capability does not mean acting without limits. External messages, public publishing, production changes, important file deletion, payment, formal submissions, sensitive accounts, and legal, financial, or medical decisions require user confirmation.

The agent can prepare content, organize material, and make suggestions, but high-risk or irreversible actions still need authorization. A useful work agent must be capable, controllable, transparent, and collaborative.

Frequently asked questions

What are M-1.5 and M1.5?

M1.5 is the intelligence core; M-1.5 is the complete system combining the model, tools, files, workflows, checks, and user controls.

How is it different from M-1?

M-1 mainly generated content. M-1.5 plans and advances multi-step work around a complete objective.

Can it perform high-risk actions automatically?

No. External and irreversible actions require explicit user confirmation.

From the first model to super agents

In February 2023, the first M-1 model and AskAIs website introduced chat and image generation. In June 2023, we opened the first API. In June 2025, M-1.5 and M1.5 brought AskAIs into a new stage.

From writing a paragraph to preparing a product launch; from answering a question to working across a multi-day startup project; from one model call to collaboration between models, tools, files, and tasks—that is M-1.5.

The AskAIs M-1.5 super agent is now in service.