Meta launched Meta Enterprise Platform on 28 September 2026, folding Muse, Meta Business Agent, and a new developer-facing Muse API under one roof, led by ex-MongoDB CEO Chirantan Desai. If you're already running a WhatsApp bot through Meta Business Agent, nothing changes today. If your team is weighing whether to build directly on Muse's API instead, the pricing has a catch worth reading before anyone signs up.
What is Meta Enterprise Platform?#
It's the new business division Meta announced this week, and it bundles four things under one roof: the Muse agent, Meta Business Agent, a developer-facing Muse API, and Muse Code, a terminal-and-CI coding agent. Per Meta's own announcement, Meta Business Agent already runs on WhatsApp, Messenger, and Instagram for over a million businesses. Chirantan "CJ" Desai (MongoDB's outgoing CEO, and a Cloudflare and ServiceNow exec before that) now reports straight to Zuckerberg as Chief Enterprise Platform Officer, and MongoDB's stock dropped over 17% the day his departure was announced. That's the market's way of saying this hire was taken seriously.
Zuckerberg's own framing: "Meta Enterprise Platform will use our strengths that few other companies have: advanced models, leading agents, large-scale infrastructure, and years of working closely with many businesses." No pricing, region list, or rollout date for the platform itself has been published yet. This is a leadership and packaging announcement, not a product launch with a signup form.
Does this change my WhatsApp bot right now?#
Short answer: no. Not if you're already on Meta Business Agent. Nothing in this announcement touches existing bot behaviour, pricing, or access for the million-plus businesses already using it through WhatsApp, Messenger, or Instagram.
Running under a couple hundred conversations a month on the free tier? This is corporate reshuffling. Skip it. If you're paying a developer or agency to build custom AI into your WhatsApp flow (a WhatsApp Business MCP setup, a lead-qualification layer, anything past the out-of-the-box agent), the new Muse API is the part that actually touches your build, because it's a cheaper new way to plug Meta's models into whatever you're already running.
What's the catch in Muse's API pricing?#
There's a real trade-off here, and it's not hidden, just easy to skim past. Muse Spark 1.3's standard tier runs $1.25 per million input tokens and $4.25 per million output, on a 1M-token context window, roughly the same range as Claude Opus 5.5. A separate Contributor tier cuts that dramatically:
| Tier | Input (per 1M tokens) | Output (per 1M tokens) | Trains on your data? |
|---|---|---|---|
| Standard | $1.25 | $4.25 | No |
| Contributor | $0.10 | $0.20 | Yes |
That's a 92-95% discount, and per the pricing listing, the reason is stated plainly: prompts and outputs on the Contributor tier "may be used to improve Meta's products." Fine for a demo. Not fine if what's running through it is client conversations, lead data, or anything under an NDA. I don't think most business owners would know to ask which tier their developer picked.
How we're handling it#
At WebEpex we build WhatsApp automation for GCC, European, US, Canadian, and Indian clients on AiSensy and self-hosted n8n rather than Meta's packaged Business Agent: we wired exactly that stack into a financial-services client's single-number setup this year, with role-based access split across roughly ten staff. That was already a control decision before this announcement.
Muse's Contributor-tier trade-off is the reason we're keeping it that way for anything client-facing. We did run one internal test against the Contributor tier this week, on a throwaway prompt set with nothing real in it, and it's genuinely fine for that: fast, cheap, no complaints. We just wouldn't point a client's data at it.
We're also not rushing anyone onto Meta's bundled business pricing on the strength of this news either, for the same reason: platform-bundled pricing moves on the vendor's schedule, not yours.
What I'd tell a client this week#
Do nothing to your existing bot. It isn't affected by any of this. If someone pitches you cheaper AI access through Muse's Contributor tier, ask what's actually flowing through it before saying yes; the standard tier is the safer default for anything touching real customer conversations. And if you're not currently paying anyone to build custom AI into WhatsApp, none of this changes your week either way.
Muse's API pricing will probably move again before the enterprise platform itself gets a public price tag. My read is Meta's still finding out what businesses will actually pay for this. If you want a second opinion on what's really running your WhatsApp setup underneath, send me what you've got and I'll tell you straight: takes two minutes, no pitch attached. [cal.com/webepex/growth-review]
