Usage pools are now live in Maximo AI Chat

We replaced the old daily message limit with two usage pools:
• 5-hour window: starts with your first request and fully resets five hours later. It does not roll forward after each request.

• Weekly pool: tracks usage across the current seven-day period and resets every seven days from your first request.

To check your limits, open a chat, open More in the sidebar, and select Usage.
You’ll see the percentage used, remaining allowance, and reset time for both pools. Use Refresh for the latest status.

Try Maximo AI: https://maximoai.co

Official X announcement: https://x.com/AskMaximoAI/status/2087902460182712460

Intelligence effort is now selectable on Maximo AI chat.

You can set reasoning effort next to the model name. Open Intelligence, pick the level, then send.

Supported levels:
• Atlas 1.2, Pandora 3.8 Preview, Pandora 3.8 Nano: Low, Medium, High, X-High, Max

• Atlas 1.1, Astra 1.7, Pandora 3.8 Pro: Low, Medium, High

If you do not pick one, it defaults to High.

Try it: https://maximoai.co

Official announcement: https://x.com/AskMaximoAI/status/2088246445426221339

New Maximo AI chat is live.

Chat now follows the actual run.

You can see reasoning and tools as they happen. The answer stays below the work. Model name and effort sit next to Maximo AI. Context usage is in the composer.

Same models. Less chrome. Easier to follow what happened.

Try it: https://maximoai.co/chat

Official announcement: https://x.com/AskMaximoAI/status/2088275899171090676

🚀 Maximo Atlas 1.2 is live

Max reasoning, built for the work that has to keep moving. Our fastest, most affordable, and most capable Atlas model yet.

What changed
The standard: staying useful after the first answer. Atlas 1.2 is tuned for sustained work that spans decisions, tool calls, corrections, and verification, across coding, research, browsing, cybersecurity, and agent workflows. This release runs at max reasoning effort.

Benchmark snapshot (max reasoning):
- 82.7% Terminal-Bench 2.1
- 82.0% BrowseComp
- 77.0% CyberGym
- 71.0% Toolathlon Verified
- 67.1% Frontier SWE
- 22.0% SWE Marathon

Who can use it
Atlas 1.2 is available now across:
- Maximo AI Chat: https://maximoai.co
- API Platform: https://maximoai.co/platform
- MyTabulon: https://mytabulon.com

Atlas 1.2 starts free in the MyTabulon Coding Plan, so developers can try it in real coding workflows before scaling up.

API pricing
Standard rates: $0.55 per 1M input, $0.05 per 1M cached input, $1.50 per 1M output tokens.

Launch offer, 80% off through August 31: $0.11 input, $0.01 cached input, $0.30 output per 1M tokens. Standard rates return September 1.

Full announcement: https://maximoai.medium.com/maximo-atlas-1-2-max-reasoning-for-work-that-has-to-keep-moving-9bfabd1b7968

Official X post: https://x.com/AskMaximoAI/status/2089301670161510486

Join the conversation in the channel and try Atlas 1.2 today.

New on Maximo AI API: /v1/responses is live

Our API now supports the OpenAI Responses API alongside Chat Completions.
What works today:
- Response creation with JSON or SSE streaming
- Stored responses (store)
- Conversation chaining with previous_response_id
- Function and tool calling
- Same auth, same USD pricing, same models
To use it, point your base URL at https://api.maximoai.co and keep your existing key.

Docs: https://maximoai.co/platform/docs#responses-api
Platform: https://maximoai.co/platform

Official announcement on X: https://x.com/AskMaximoAI/status/2094752770125414452

Maximo Atlas 1.3 is here

Our most capable Atlas model yet. A full-stack frontier agent built for long-horizon terminal, code, and web work, tuned for max reasoning effort. Not high. Not xhigh. Max.

What it delivers:
• 84.4% on Terminal-Bench 2.1, up from 82.7% on Atlas 1.2
• 39.0% on Terminal-Bench 4.0, the hardest terminal tier
• 64.0% on DeepSWE 1.1, up from 55.7%
• 71.0% on Toolathlon Verified, holding the tool-reliability bar
• 55.7% on Humanity's Last Exam with tools, 43.4% without
• 1774 Elo on GDPval-AA v2, up from 1706
• 85.7% on BrowseComp, up from 82.0%
• 77.0% on CyberGym

Pricing built for work that runs all day:
• Input: $0.20 per 1M tokens
• Cached input: $0.02 per 1M tokens
• Output: $0.50 per 1M tokens

Cached input at $0.02 is where long agent sessions win, since the same context gets read many times in one run.

Who it is for:
Developers building coding agents, teams researching unfamiliar codebases and markets, operators automating browser work, security practitioners, and founders turning ideas into working prototypes.

Try it now:
• Maximo AI Chat: https://maximoai.co
• Maximo AI API Platform: https://maximoai.co/platform
Model ID on the API: maximo-atlas-1.3

Full benchmark matrix, methodology, and pricing breakdown in the launch article:
https://maximoai.medium.com/maximo-atlas-1-3-the-full-stack-frontier-agent-priced-like-a-workhorse-2b23c0d156d7

Also announced on X: https://x.com/AskMaximoAI/status/2095111651917791458