Akili

Real numbers, not a quote request.

Pricing

A plan, plus credits that scale with how much you actually use it. Every tier below is a real, live price — not a placeholder waiting on a sales call.

$1 · 14-day Pro trial

Try everything in Pro for 14 days for $1 — 500 credits, every model provider, agents, knowledge bases, and MCP servers. Cancel in one click before day 14 and you've spent a dollar. Keep it and you roll onto Pro.

Start the $1 trial

Individual

Free

Get started

$0

500 credits/mo

Use for free
  • 25 messages per hour
  • 3 knowledge bases, 50 docs each
  • 5 custom agents
  • 1 webchat widget
  • Data export

Pro

Most popular

Research, code, and organize

$18/mo

$15/mo billed annually

1,000 credits/mo

Try Pro — $1 for 14 days
  • Server-side AI models included
  • 100 messages/hr, 500/day
  • MCP servers & code execution
  • Content Studio access
  • Everything in Free

Max

Higher limits, priority access

$100/mo

$85/mo billed annually

8,000 credits/mo

Get Max
  • Up to 20× more usage than Pro
  • All server-side AI models
  • SMS two-factor authentication
  • Priority processing
  • Everything in Pro

Team & organization

Team

For growing teams

$30/seat/mo

$25/seat/mo billed annually

5–50 seats

12,000 credits/mo, pooled org-wide

Get Team
  • Organization workspace
  • Pooled org-wide credit allowance
  • Slack & Teams deployment (5 each)
  • Shared knowledge bases
  • Marketplace publishing

Business

Most popular

Enterprise-grade security & scale

$55/seat/mo

$45/seat/mo billed annually

20–2,000 seats

60,000 credits/mo, pooled org-wide

Get Business
  • Enterprise SSO (SAML/OIDC)
  • SCIM deprovisioning
  • Immutable audit logs
  • Require 2FA enforcement
  • Priority support

Enterprise

Custom deployment & SLAs

Custom

Unlimited seats

Negotiated credit pool

Talk to us
  • Self-hosted deployment support
  • Custom SLA guarantees
  • Dedicated support channel
  • Early access AI models
  • Everything in Business

Credits, explained

One unit, every model

A credit covers roughly 1,000 tokens, cost-anchored to real model cost. Heavier models draw credits faster — nothing else changes.

Pooled for teams

Team and Business pools are org-wide: your researcher can have a heavy week while the rest of the team stays light.

No surprise overage

Usage pauses at zero by default. Auto-refill is opt-in and capped at a limit your admin sets — never an open-ended bill.

The full mechanics — allowances, top-up packs, roll-over — are in Credits & billing.

Need something custom?

Enterprise covers self-hosted deployment, custom SLAs, unlimited seats, and dedicated support. It's genuinely custom — that's why it's the one tier we can't put a number on.

Talk to us about your environment

Pricing questions, answered directly

One dollar gets you 14 days of Pro — every model provider, custom agents, knowledge bases, MCP servers, and 500 credits to spend. Cancel any time before day 14 and you've spent $1 total. Do nothing and you roll onto Pro at $18/mo (or $15/mo billed annually).

A credit is Akili's usage unit — roughly 1,000 tokens of model usage, cost-anchored so a heavier model simply consumes credits faster. Every plan includes a monthly credit allowance; included credits reset monthly, purchased top-up credits roll over.

Usage pauses by default once your pool is spent for the month. You can turn on a capped auto-refill yourself if you'd rather not think about it — there's no surprise overage charge either way.

Both, on purpose. Team and Business are priced per seat, and each plan includes a pooled, org-wide credit allowance everyone draws from. It's not pure per-seat, and it's not pure metered usage — it's built to match how teams actually use an AI platform.

No. With bring-your-own-key, model usage bills directly to your provider account with no markup from us. Included server-side model usage draws from your credit allowance instead.

Yes — upgrade or downgrade at any time. Team and Business plans scale seats up as you grow, with no re-negotiation required.

Self-hosted deployments are licensed per named user and scoped to your environment — you license the software, not the tokens, because inference runs on your own keys. It's genuinely custom to your infrastructure, so we don't publish a number: talk to us.