Language model · from 1 credit
Write and analyze with Mercury 3
Mercury 3 is Inception's new diffusion model, released on October 19, 2026, and the successor to Mercury 2.5. Diffusion refines every word of an answer in parallel instead of writing them one by one: in our tests, a 150-word pitch came out in under two seconds. The context grows to one million tokens and the price stays among the lowest in the catalogue, slightly above Mercury 2.5. It reads text only: images connected to the node are not sent to it, and the node says so. It is the model for repetitive work in bulk, rephrasing, summarizing, spinning a text into versions, when speed matters more than finesse.
A model that writes everything at once
Mercury 3 was released on October 19, 2026 and takes over from Mercury 2.5. Like it, it is a diffusion model: instead of writing its answer word by word, it starts from a complete draft and refines it in a few passes. You notice the speed: in our tests on October 21, 2026, a one-word answer arrived in 0.7 seconds and a 150-word pitch in 1.6 seconds.
Compared with Mercury 2.5, the context grows from 260,000 to one million tokens, and Inception presents it as its most capable model for multi-step tasks. The price goes up a little and stays among the lowest in the catalogue.
What it does well on the canvas
Mercury 3 is for repetitive tasks where speed matters: rephrasing an instruction, summarizing a connected text, spinning one prompt into ten variations for Image nodes downstream. For a long script or a fine analysis, Claude Sonnet 5.5 or Gemini 4 Argon remain safer.
It reads text only. If an image is connected to the Text node, it is not sent to it and the node tells you: pick a model that reads images, Claude Haiku 5.5 for example, to describe a render.
Mercury 3 next to its closest siblings
Figures pulled from the same registry as billing: when a price or a cap changes, this table changes with it.
| Mercury 3 | Claude Haiku 5.5 | DeepSeek V3.2 | |
|---|---|---|---|
| Base price | 1 credit(s) / message | 1 credit(s) / message | 1 credit(s) / message |
How to use it on Imaginode
- 1Add a Text node and pick Mercury 3 in the selector.
- 2Give your instruction; wire in other texts or images to analyze.
- 3The result feeds your connected Image and Video nodes directly.
Frequently asked questions
Does Mercury 3 replace Mercury 2.5?
Both remain available. Mercury 2.5 is even cheaper for simple tasks; Mercury 3 handles long instructions and large documents better.
Why are my images not analyzed?
Mercury 3 accepts text only: the provider rejects any image. The node drops them before sending to avoid a failure, and warns you.
How much does a message sent to Mercury 3 cost?
Around 1 credit(s) per message, that is $0.01 (1 credit = $0.012). The cost shows before sending, with no separate subscription and no API key to manage: the same credits pay for all 129 models in the catalog.
Do I need a separate subscription to use Mercury 3?
No: Mercury 3 is included in Imaginode, with no API key and no account with the provider. A single credit balance covers all 129 models, text, image, video and voice, and also powers the assistant that builds complete workflows.
Can I try Mercury 3 for free?
Yes: signing up grants trial credits, no credit card required. Enough to test Mercury 3 in real conditions before deciding on a plan or a top-up.
Do I need to install anything to use Mercury 3?
No: Mercury 3 runs server-side, in the browser, mobile included. Imaginode's node canvas wires its renders to the other models (references, camera, styles) with nothing to install.
Try Mercury 3 on a real canvas
129 models, one account, credits that work everywhere. The price always shows before you generate.
Get started
GPT-5 (ChatGPT)