ChatImage

Model developer guide

Meta

5 website model guides

Choose Muse Wonder for demanding coding and reasoning, Muse Spark for established chat workflows, or Muse Image for visual creation. Wonder and Spark return text; Muse Image creates or edits images. Read the exact model guide for input, reasoning, data-tier and app integration details.

ReasoningCodingMultimodal understanding

Practical guidance

Where should I start?

Editorial starting points, not a performance ranking. Check current app access before choosing a model.

Work through a demanding code or analysis brief

Evaluate Meta’s preview muse-wonder-1.0 for complex coding and multi-step reasoning. Supply clear source material and ask for a bounded deliverable with evidence and verification steps.

Tradeoff: Wonder Full is a different Muse Code variant. Higher reasoning effort stays on Wonder rather than selecting Full, and more reasoning does not guarantee correct code or complete recall.

Muse Wonder guide

Understand a complex brief or code change

Meta recommends Muse Spark 1.3 for new coding and multi-step work. Start with a bounded document or software question and ask for a text answer tied to evidence. Its multimodal input support does not mean it generates images or spoken audio.

Tradeoff: The provider warns that audio understanding in 1.3 is not fully supported and may be degraded. Do not choose it for an audio-critical task merely because the modality appears in a table.

Muse Spark 1.3 guide

Create or revise a visual

Muse Image is a separate family for text-led image creation and reference-image editing. Use an explicit visual brief or a carefully scoped revision; the model overview distinguishes image output from Muse Spark’s text output.

Tradeoff: A provider feature such as conversational editing or built-in search does not establish that the app exposes its controls or preserves every prior reference across turns.

Muse Image 1.0 guide

Technical details

What does the provider document?

Official documentation describes the named model or gateway—not every feature available in this app.

Muse Spark shares context, not identical behavior

The model table lists a 1,048,576-token context for Spark 1.1, 1.2 and 1.3, with text, image, video, audio and PDF input and text output. The 1.3 audio caveat qualifies that table. Context capacity is separate from file ingestion limits and does not guarantee that every detail is recalled.

Reasoning effort and data tier are linked

Muse Spark does not accept reasoning_effort none. The max level is restricted to Standard-tier Muse Spark 1.3; it is not available on Contributor variants or the earlier Spark versions. Meta describes Contributor as a discount in exchange for permission to train on prompts and completions, so it is not just another speed setting.

Muse Image has distinct endpoint options

Meta documents Responses for conversational editing and separate images/generations and images/edits endpoints for single-shot work. The model is muse-image-1.0. These provider routes do not establish which tools, references or continuity settings the app forwards; the audited app route is described separately below.

Complete website coverage

Explore all 5 model guides

Every existing catalog entry for this provider is linked below. Status labels come from the individual website guides; catalog coverage is not a guarantee of current app access. Historical and snapshot pages remain references, not extra active models.

Practical guidance

From a brief to a useful result

Original prompt examples, not generated-output demonstrations. They have not been generation-tested.

Workflow 1

Find contradictions in a project handover

Provide labeled excerpts from a plan, status note and acceptance checklist. Use text first; attach visuals only if your app session explicitly supports them.

Compare the three handover documents below. Identify conflicting dates, missing owners and requirements that appear in only one document. Cite the document label and exact supporting phrase for each issue. Produce a short reconciliation checklist, separating facts we can settle from the text from questions that need the project owner. Do not invent missing commitments.

Check the result: Verify every quoted phrase and confirm that a later decision was not mistaken for a contradiction. Assign unresolved questions to a person rather than accepting a guessed resolution.

Use the Muse Spark 1.3 guide

Workflow 2

Design a clear event illustration

Use a simple original brief with one focal point. Keep essential event details as editable website text rather than relying on text baked into an image.

Create an original editorial illustration for a neighborhood repair workshop: two adults repairing a small wooden chair at a tidy workbench, with a few clearly arranged hand tools. Use warm daylight, an approachable color palette and ample uncluttered space above the scene. Avoid logos, readable lettering, extra fingers and dramatic effects. The image should support a practical community event, not look like a product advertisement.

Check the result: Inspect hands, tool shapes, safety details and unwanted lettering. Confirm the crop leaves space for accessible HTML event information and does not communicate an unintended brand claim.

Use the Muse Image guide

Technical details

What changes inside EZ Ai Assist?

Official Meta documentation supports the provider capabilities below, including Wonder’s model identity, preview status and reasoning controls. Muse Image integration details retain their separate October 8 owner-audit attribution. Earlier Spark integration details remain unverified; provider references are not live generation tests.

Website catalog

Which Meta entries are represented?

This site has Muse Wonder, three Muse Spark version guides and one Muse Image guide. It does not add every family in Meta’s documentation to the app catalog. Separate Meta-developed models hosted by another gateway keep their existing website classification.

Owner audit · October 8

What is reported for image generation?

The existing Muse Image guide reports app ID muse-image-1-0, upstream muse-image-1.0 through Meta Model API, the Responses route and WebP output. These are owner-audit statements, not fresh live-generation evidence or a guarantee that all direct endpoint options are exposed.

Meta model and reasoning documentation

Keep Wonder and Wonder Full distinct

The Wonder guide covers muse-wonder-1.0 on Meta Model API. Chat Completions and Responses use their respective reasoning fields, while Responses provides web search with citations. Choosing max effort does not select Wonder Full or transfer Muse Code subscription terms to the API model.

Before you rely on a result

Limits, costs and tradeoffs

  • Understanding an image, audio clip or document is not the same as reproducing its content perfectly. Check source readability and request explicit uncertainty where evidence is unclear.
  • Muse Spark is hosted on Meta Model API. Do not borrow licensing, self-hosting or data promises from Llama, Muse Glimmer or another Meta family.
  • Meta’s provider token, image and tool charges are separate accounting units. Published rates and tier labels are not a quote for an EZ Ai Assist task.

Provider references and picker estimates are not EZ Ai Assist customer charges. Read the website pricing explanation.

Compare by task

Other paths to explore

Google

Compare another multimodal family while keeping input analysis separate from media generation.

OpenAI

Explore paired text and image workflows with different model-specific contracts.

FLUX

Consider an image-focused family when the deliverable is a visual rather than a text analysis.

Common questions

Meta, without the assumptions

Is Muse Spark another name for Llama?

No. This guide covers Meta’s hosted Muse Spark family. A shared developer does not make different model families, licenses or serving routes interchangeable.

Can Spark 1.3 reliably analyze an audio recording here?

The provider warns its audio support is incomplete and recommends other options for audio. App audio ingestion is also unverified, so this guide does not promise that workflow.

Does selecting Meta mean my prompts are used for training?

The cited provider documentation distinguishes Standard and Contributor terms. The actual app tier was not inspected, so neither a training-use claim nor a no-training guarantee is established here. Confirm the route-specific terms before submitting sensitive material.

New to the terminology?

Official documentation

Sources & review notes

Highlighted provider contracts reviewed . Individual model guides carry their own review dates. Sources can change; these pages do not perform live health checks.

Official Meta documentation supports the provider capabilities below, including Wonder’s model identity, preview status and reasoning controls. Muse Image integration details retain their separate October 8 owner-audit attribution. Earlier Spark integration details remain unverified; provider references are not live generation tests.