Image input limitUp to 16 images; each up to 10 MiB. PNG, JPEG, WebP or GIF in base64 data URLs. No remote URLs or images in criteria.See specification notes below
Discovery metadata reports this context value; the decision guide does not independently document a token limit.
Maximum input tokens
Not collected
No exact model-specific contract established in the reviewed evidence.
Maximum output
Not collected
Discovery lists 50,000, but the decision request has no generation-budget parameter. This generic value is not treated as a configurable output allowance.
Parameters
Not collected
No exact model-specific contract established in the reviewed evidence.
Release date
Not collected
No exact model-specific contract established in the reviewed evidence.
Knowledge cutoff
Not collected
No exact model-specific contract established in the reviewed evidence.
Reasoning
Not collected
No exact model-specific contract established in the reviewed evidence.
Prices retain their published currencies and billing units below.
Published billing units · original currency · observed 15 Oct 2026
Category
Usage
Rate
Unit
Typed decision tokens
Input tokens1
$0.02
1M tokens
Typed decision tokens
Output tokens1
$0.00
1M tokens
Typed decision tokens
Cached input1
$0.00
1M tokens
Conflicting discovery price
Input tokens2
$0.04
1M tokens
Rate conditions
Documentation advertises a 50% input promotion without an end date, but web-fetched discovery lists $0.04/1M input (the direct fetch lists $0.02). Confirm account billing. Image requests and multiple named questions are not equivalent to a standard chat question.
Web-fetched discovery metadata differs from the $0.02/1M promotional documentation. Account-specific billing is unverified.
Terms permit using submissions for model training unless the user disables “Improve the model for everyone” in API Platform settings. The privacy policy uses purpose-dependent retention, not a fixed universal period. The FAQ says custom zero-retention arrangements are handled case by case; enterprise agreements may differ.
Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.
01 / INITIAL INPUTProcess & retain
The system processes eligible input and retains reusable state.
02 / LATER REQUESTCheck available state
Reuse depends on matching input, session state and the cache policy.
03 / CACHE HITReuse & generate
Reusing state can reduce repeated work. Any charges follow the billing rules below.
Inception API
mercury-decide
Not verified
No version-specific prompt-cache contract established.
How it works
Not collected
No version-specific prompt-cache contract established.
Minimum reusable input
Not collected
No version-specific prompt-cache contract established.
Lifetime & refresh
Not collected
No version-specific prompt-cache contract established.
How to enable it
Not collected
No version-specific prompt-cache contract established.
Usage fields
Not collected
No version-specific prompt-cache contract established.
Billing rules
Discovery lists zero cached-input and cache-write prices; decision usage has no cached-token counter. A usable prefix-cache contract is not established.Source for Billing rules (opens in a new tab)
Limits & data boundaries
Not collected
No version-specific prompt-cache contract established.
Documentation advertises a 50% input promotion without an end date, but web-fetched discovery lists $0.04/1M input (the direct fetch lists $0.02). Confirm account billing. Image requests and multiple named questions are not equivalent to a standard chat question.
Observed 15 Oct 2026. Write rates and storage fees are different charges; use the billing rules above.
Caching reviewed 15 Oct 2026
What the measurements say
No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.
Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.
The record behind the model
Why it is included
Inception’s own Trust Center describes the Mercury family as proprietary. Public API and enterprise deployment access do not establish public weights.
