Glean Enterprise Flex is the pricing model Glean describes as combining the budgeting simplicity of a per-user licence with a pooled allowance of pay-per-use credits, so that an organisation pays only when it goes above the initial allocation.[1] Glean Core Suite is the alternative seat-based model, in which the seat covers the platform and model use is bought separately as Model Hub Usage.[2] This article explains how each model counts users and AI consumption. The overview is Glean licensing.
Editions
Glean Enterprise Flex
The Enterprise Flex Seat is licensed per user per month and is designed to be deployed to every employee, offering extensive use of everyday AI search and chat across the organisation’s connected data.[1] The documentation divides capabilities into those included in the seat and those that consume FlexCredits.
| Capability | Treatment under Enterprise Flex[1] |
|---|---|
| Basic search | Unlimited |
| Fast Mode Assistant queries | Unlimited |
| Thinking and Adaptive Reasoning queries, standard models | Included up to 100 per user per week; excess consumes FlexCredits |
| Thinking and Adaptive Reasoning queries, premium models | Consume FlexCredits |
| Code Writer, slide generation, meeting notes | Consume FlexCredits |
| Deep Research | On the Glean Universal Key, billed as an Assistant query; on a Customer Key, consumes FlexCredits per run until it becomes a skill |
| Agent creation, testing and sharing | Unlimited |
| Agent runs | Consume FlexCredits |
| Collections, go links and the people directory | Unlimited |
| Client APIs (Search, Chat, Agents) and the MCP server | Consume FlexCredits |
| Platform, admin, indexing, connector and tools APIs | Unlimited |
| Memory | Saved memories included in the seat; learned memories consume FlexCredits |
Embedded integrations in Slack, Zoom, Microsoft Teams, GitHub and Miro consume Fast Mode queries, Thinking queries and agent runs at the same entitlements and rates as the Assistant.[1] Glean Protect+ and Premium Support are each licensed site-wide as an annual fee.[1]
Glean Core Suite
Core Suite is licensed on a per-user, per-month basis. The seat fee includes the Enterprise Graph, hybrid search index, agent harness, connectors with native read and write tools, an MCP host, Glean Protect controls, billing and insights dashboards, search, the library, agent building and governance, the browser, desktop and mobile surfaces, and platform and admin APIs. Glean Hosting is included in the seat fee for Glean Hosted customers.[2] Features that call a model, such as reasoning modes, canvas, code writer, deep research, meeting notes, voice, the Chat API and the Agents API, draw Model Hub Usage.[2]
Legacy Glean Enterprise
A customer on a legacy Glean Enterprise agreement receives some Enterprise Flex capabilities at Glean’s discretion for the length of its current term and then must move entirely to Enterprise Flex. At that point it receives an allocation of base credits with the seat, and earlier FlexCredit balances do not entitle it to Flex features.[1]
Metrics
FlexCredits
FlexCredits are shared and pooled at the organisation level. Enterprise Flex Agreements always include a base pool, and more can be added at any time by purchasing additional FlexCredit Packs.[1] Metered capabilities consume credits at runtime, either on a fixed basis per run or on a variable basis depending on the complexity of the task. Glean describes a FlexCredit as a measure of that complexity, including agentic planning, data indexed and retrieved, response synthesis, tools used and the underlying compute and model infrastructure.[1] The rate card gives typical consumption at the 50th and 90th percentile.
| Capability | Unit | Typical FlexCredits, 50th percentile | 90th percentile |
|---|---|---|---|
| Fast Mode Assistant query | Query | about 3 | about 15 |
| Thinking Mode query, basic or standard models | Query | about 7 | about 26 |
| Thinking Mode query, premium models | Query | about 35 | about 120 |
| Adaptive Reasoning query, standard models | Query | about 11 | about 38 |
| Adaptive Reasoning query, premium models | Query | about 26 | about 83 |
| Meeting notes | Meeting | about 9 | about 18 |
| Code Writer query | Query | about 9 | about 32 |
| Slide generation | Query | about 45 | about 142 |
| Glean Agent Run | Run | about 7 | about 114 |
| Memory and personalisation | Monthly active user | about 5 | about 10 |
Fast Mode queries carry a rate of 0 and are unlimited under the seat, so the first row shows what the capability uses and does not bill. The figures are described as typical and not as list rates, and the rate card is not fixed.[1] A Thinking query that consumes more than 30 FlexCredits is counted as several queries against the weekly allowance, and a follow-up prompt counts as an additional query.[1] Learned memories are free through 6 November 2026 and begin to bill on 7 November 2026.[1]
Model Hub Usage
Model Hub Usage is an upfront commit for large language model usage, charged at model or hosting provider API rates. Glean’s own models are charged at listed rates, and for customers using a Customer Key it is limited to the Glean Model Family.[2] The page lists rates per million tokens by model, for example 0.50 US dollars input and 2.50 output for Glean’s Waldo model, with each provider’s rate shown for its models. Glean charges only for successful LLM calls, meaning a request sent to a model that completes successfully.[2] Because the rates follow each provider’s API pricing, the cost of a given user action changes whenever a provider changes its prices or when a model is replaced.
Model tiers
Models are categorised as Basic, Standard or Premium by reasoning capability and cost. Basic and Standard models in Fast Mode are included in Enterprise Flex Seats, while premium models consume FlexCredits. When the auto model selector picks a premium model, FlexCredits are consumed, and all premium models can be disabled in the Admin console.[1] Model tiers are subject to change, and some premium models are available only to customers using a Customer Key.[1]
Counting / floors
- Included allowance. Thinking and Adaptive Reasoning queries with standard models are included up to 100 per user per week.[1]
- Pool and packs. The FlexCredit pool is organisational, so one department’s heavy use reduces what is left for others unless limits are set. See Glean usage limits, connectors and agents.
- API coverage. On Core Suite, API usage included in the seat covers normal human-scale product and portal usage, and machine-scale or abusive volumes without corresponding seat coverage are subject to additional fees.[2]
- Own keys and self-hosting. Enterprise Flex Seats and FlexCredits are discounted for customers that supply their own LLM keys or self-host Glean in their private cloud, to compensate for compute and model costs they bear.[1]
- Add-ons. Flexible Model Management is charged as a percentage of total LLM usage, based on upfront commit or total usage, and includes unified model management, traffic performance management, evals, granular model choice and intelligent routing.[2]
Virtualization & partitioning
Not applicable to processors. A seat is per user. Departments, which the Usage dashboard attributes from identity-provider metadata, are a reporting partition and a place to set limits, and moving a user between departments changes how usage is attributed.[3]
Cloud / BYOL
Glean can run in the customer’s cloud account or in Glean’s. A Customer Key organisation keeps a different Deep Research and agent-run treatment than a Glean Universal Key organisation, so the model key choice changes billing as well as retention commitments.[1]
Programs
- FlexCredit Pack. Additional credits for Enterprise Flex organisations.[1]
- Glean Flexible Model Management. Percentage-of-usage fee for model governance and routing.[2]
- Glean Protect+. Site-wide annual licence for sensitive content models and prompt-injection protection.[1]
- Legacy transition. Move from Glean Enterprise to Enterprise Flex at term end.[1]
Out of scope
- Seat prices, FlexCredit pack prices and discounts, which Glean does not publish.
- The full per-model token rate table, which changes as providers revise their prices.
- Customers’ separate agreements with model providers where they use their own key.