Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Meshy / Meshy Smart Topology

Meshy T2

Native mesh generation with controllable face count and separated parts; available through Smart Topology API workflows.

3D generation
availableClosed weights3D generation
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
11 of 25 specification fields collected
3D output formatsGLB, OBJ, FBX, STL, USDZ, 3MF (opt-in)See specification notes below
Text input limit800800
Input billingSee ratesMeshy API · text preview
Output billingSee ratesPublished units and conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected
Maximum input tokens
Not collected
Maximum output
Not collected
Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected

API task JSON is not caller-defined structured model output.

Streaming
Not collected

SSE reports task progress; streamed geometry output is not established.

Batch processing
Not collected

Multi-view inputs and queued tasks do not establish a discounted batch API.

Fine-tuning
Not collected

No customer model-weight fine-tuning contract documented.

Knowledge cutoff
Not collected
Release date
Not collected

Original release date not established by a later API or research announcement.

Parameters
Not collected
Architecture
Flow matching over continuous vertex tokens with parallel mesh generation.Source for Architecture (opens in a new tab)
Languages
Not collected
Context caching
Not collected

No documented cache controls or cache price.

Output modalities
3D meshSource for Output modalities (opens in a new tab)
Input modalities
Text, ImageSource for Input modalities (opens in a new tab)

Model-specific operation contracts below distinguish direct inputs from texture guidance.

Maximum prompt characters
800Source for Maximum prompt characters (opens in a new tab)

Text-to-3D prompt limit, not tokens.

Image input limits
One image; multi-view is a separate operation where supported.Source for Image input limits (opens in a new tab)
Input image formats
JPG, JPEG, PNG, WebPSource for Input image formats (opens in a new tab)
Maximum input content size
20,000,000Source for Maximum input content size (opens in a new tab)

Provider states 20 MB per image; byte value uses decimal MB.

Output formats
GLB, OBJ, FBX, STL, USDZ, 3MF (opt-in)Source for Output formats (opens in a new tab)

3MF must be explicitly requested. Geometry and texture assets differ by task.

Usage metering
Prepaid API credits per task; geometry and texture stages may be billed separately.Source for Usage metering (opens in a new tab)
Mesh topology and face count
Triangle-only native mesh; 100–15,000 target faces, default 4,000.Source for Mesh topology and face count (opens in a new tab)
Endpoint availability
Hosted model is available; official repository still describes an open-source release as forthcoming.Source for Endpoint availability (opens in a new tab)

Reviewed tree contains README and teaser only; no model weights or inference code. Recheck before assuming permanent closed status.

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Meshy API · text preview

Provider source (opens in a new tab)
API model ID
meshy-t2

Service selector, not an immutable weight hash.

Access
POST https://api.meshy.ai/openapi/v2/text-to-3d with Bearer API key and ai_model=meshy-t2. Use mode=preview. Set model_type=smart-topology. Task polling, listing, deletion and SSE status endpoints are documented; no inference was performed.
Regions
No verified processing region
Asynchronous generation
SupportedSource for Asynchronous generation (opens in a new tab)

Background tasks with polling or SSE status updates; no progressive mesh streaming is inferred.

Data retention
API output is deleted after three days for non-Enterprise customers. Enterprise defaults to indefinite output storage, configurable to 1–30 days or contract-specific terms. This is output retention, not a guarantee about logs or all customer data.Source for Data retention (opens in a new tab)
Request quotas
Account-wide: Pro/Premium/Ultra/Studio 20 requests/sec; queue limits 10/30/100/20 respectively. Enterprise 100 requests/sec and default queue 100, customizable. Web tasks share the queue.Source for Request quotas (opens in a new tab)
Cancellation and task deletion
Pending deletion refunds credits. In-progress deletion returns 409 and leaves generation running. Completed/failed/canceled tasks can be deleted.Source for Cancellation and task deletion (opens in a new tab)

Behavior documented September 11, 2026. No paid task was submitted.

Output transport
Completed file URLs and task metadata; expires_at is a millisecond timestamp.Source for Output transport (opens in a new tab)
Input modalities
TextSource for Input modalities (opens in a new tab)
Output modalities
3D meshSource for Output modalities (opens in a new tab)
Output formats
GLB, OBJ, FBX, STL, USDZ, 3MF (opt-in)Source for Output formats (opens in a new tab)

Explicit target_formats controls output. The default omits 3MF.

Geometry resolution
No Ultra geometry option documented for this model.Source for Geometry resolution (opens in a new tab)
Mesh topology and face count
Triangle only; target faces 100–15,000, default 4,000. Remeshing and adaptive decimation ignored.Source for Mesh topology and face count (opens in a new tab)
Input moderation
Optional input moderation; default false.Source for Input moderation (opens in a new tab)
Character pose
a-pose, t-pose or empty string; defaults to empty.Source for Character pose (opens in a new tab)
Maximum prompt characters
800Source for Maximum prompt characters (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
3D generationMesh only15 creditsrequest

Rate conditions

  1. Preview geometry only. A refine request is billed separately.
Pricing source (opens in a new tab)

Data & enterprise terms

API output is deleted after three days for non-Enterprise customers. Enterprise defaults to indefinite output storage, configurable to 1–30 days or contract-specific terms. This is output retention, not a guarantee about logs or all customer data.

Read the service policy (opens in a new tab)

Meshy API · image to 3d

Provider source (opens in a new tab)
API model ID
meshy-t2

Service selector, not an immutable weight hash.

Access
POST https://api.meshy.ai/openapi/v1/image-to-3d with Bearer API key and ai_model=meshy-t2. Set model_type=smart-topology. Task polling, listing, deletion and SSE status endpoints are documented; no inference was performed.
Regions
No verified processing region
Asynchronous generation
SupportedSource for Asynchronous generation (opens in a new tab)

Background tasks with polling or SSE status updates; no progressive mesh streaming is inferred.

Data retention
API output is deleted after three days for non-Enterprise customers. Enterprise defaults to indefinite output storage, configurable to 1–30 days or contract-specific terms. This is output retention, not a guarantee about logs or all customer data.Source for Data retention (opens in a new tab)
Request quotas
Account-wide: Pro/Premium/Ultra/Studio 20 requests/sec; queue limits 10/30/100/20 respectively. Enterprise 100 requests/sec and default queue 100, customizable. Web tasks share the queue.Source for Request quotas (opens in a new tab)
Cancellation and task deletion
Pending deletion refunds credits. In-progress deletion returns 409 and leaves generation running. Completed/failed/canceled tasks can be deleted.Source for Cancellation and task deletion (opens in a new tab)

Behavior documented September 11, 2026. No paid task was submitted.

Output transport
Completed file URLs and task metadata; expires_at is a millisecond timestamp.Source for Output transport (opens in a new tab)
Input modalities
ImageSource for Input modalities (opens in a new tab)
Output modalities
3D meshSource for Output modalities (opens in a new tab)
Output formats
GLB, OBJ, FBX, STL, USDZ, 3MF (opt-in)Source for Output formats (opens in a new tab)

Explicit target_formats controls output. The default omits 3MF.

Geometry resolution
No Ultra geometry option documented for this model.Source for Geometry resolution (opens in a new tab)
Mesh topology and face count
Triangle only; target faces 100–15,000, default 4,000. Remeshing and adaptive decimation ignored.Source for Mesh topology and face count (opens in a new tab)
Input moderation
Optional input moderation; default false.Source for Input moderation (opens in a new tab)
Character pose
a-pose, t-pose or empty string; defaults to empty.Source for Character pose (opens in a new tab)
Input image formats
JPG, JPEG, PNG, WebPSource for Input image formats (opens in a new tab)
Image input limits
One image; maximum 20 MB.Source for Image input limits (opens in a new tab)
Required request fields
image_url or completed single-image input_task_id; task ID takes precedence.Source for Required request fields (opens in a new tab)
Texture resolution
2K, 4K or 8K base color; default 2KSource for Texture resolution (opens in a new tab)
PBR materials
Optional metallic, roughness and normal maps. No emission map.Source for PBR materials (opens in a new tab)

At 8K base color, PBR maps are 4K per July 21 changelog.

Input image enhancement
Unsupported; providing image_enhancement returns an error.Source for Input image enhancement (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
3D generationMesh only15 creditsrequest
3D generationMesh with 2K texture215 creditsrequest
3D generationMesh with 4K texture315 creditsrequest
3D generationMesh with 8K texture420 creditsrequest

Rate conditions

  1. Set should_texture=false; no texture phase.
  2. Complete task rate including geometry and selected texture; not an extra charge on top of the mesh-only row.
  3. Complete task rate, including texture.
  4. Complete task rate; PBR at 4K and no emission.
Pricing source (opens in a new tab)

Data & enterprise terms

API output is deleted after three days for non-Enterprise customers. Enterprise defaults to indefinite output storage, configurable to 1–30 days or contract-specific terms. This is output retention, not a guarantee about logs or all customer data.

Read the service policy (opens in a new tab)

Meshy API · text refine

Provider source (opens in a new tab)
API model ID
Not collected

No immutable public selector verified for this historical access record.

Access
POST https://api.meshy.ai/openapi/v2/text-to-3d with mode=refine and a completed preview_task_id. Omit ai_model to inherit the T2 preview; explicit meshy-t2 is not in the refine selector list.
Regions
No verified processing region
Asynchronous generation
SupportedSource for Asynchronous generation (opens in a new tab)

Background tasks with polling or SSE status updates; no progressive mesh streaming is inferred.

Data retention
API output is deleted after three days for non-Enterprise customers. Enterprise defaults to indefinite output storage, configurable to 1–30 days or contract-specific terms. This is output retention, not a guarantee about logs or all customer data.Source for Data retention (opens in a new tab)
Request quotas
Account-wide: Pro/Premium/Ultra/Studio 20 requests/sec; queue limits 10/30/100/20 respectively. Enterprise 100 requests/sec and default queue 100, customizable. Web tasks share the queue.Source for Request quotas (opens in a new tab)
Cancellation and task deletion
Pending deletion refunds credits. In-progress deletion returns 409 and leaves generation running. Completed/failed/canceled tasks can be deleted.Source for Cancellation and task deletion (opens in a new tab)

Behavior documented September 11, 2026. No paid task was submitted.

Output transport
Completed file URLs and task metadata; expires_at is a millisecond timestamp.Source for Output transport (opens in a new tab)
Maximum prompt characters
800Source for Maximum prompt characters (opens in a new tab)

Optional texture_prompt, not a new geometry prompt.

Input modalities
Preview task, Text or image texture guidanceSource for Input modalities (opens in a new tab)
Output modalities
3D meshSource for Output modalities (opens in a new tab)
Output formats
GLB, OBJ, FBX, STL, USDZ, 3MF (opt-in)Source for Output formats (opens in a new tab)

Explicit target_formats controls output. The default omits 3MF.

Texture resolution
Inherited T2 refine: no independent higher-resolution contract verified. Explicit texture-model override can change capabilities.Source for Texture resolution (opens in a new tab)
Required request fields
mode=refine and preview_task_id required. texture_prompt takes priority over texture_image_url if both are supplied.Source for Required request fields (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
3D generationRefine texture at 2K110 creditsrequest

Rate conditions

  1. Texture generation only; preview paid separately.
Pricing source (opens in a new tab)

Data & enterprise terms

API output is deleted after three days for non-Enterprise customers. Enterprise defaults to indefinite output storage, configurable to 1–30 days or contract-specific terms. This is output retention, not a guarantee about logs or all customer data.

Read the service policy (opens in a new tab)

Caching

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.

Meshy API · text preview

meshy-t2
Not verified

No model-specific prompt-cache contract or discounted cache rate documented.

How it works
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Minimum reusable input
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Lifetime & refresh
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

How to enable it
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Usage fields
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Billing rules
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Limits & data boundaries
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

Meshy API · image to 3d

meshy-t2
Not verified

No model-specific prompt-cache contract or discounted cache rate documented.

How it works
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Minimum reusable input
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Lifetime & refresh
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

How to enable it
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Usage fields
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Billing rules
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Limits & data boundaries
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

Meshy API · text refine

Not verified

No model-specific prompt-cache contract or discounted cache rate documented.

How it works
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Minimum reusable input
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Lifetime & refresh
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

How to enable it
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Usage fields
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Billing rules
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

Limits & data boundaries
Not collected

No model-specific prompt-cache contract or discounted cache rate documented.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 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

The official model repository says its open-source release is being prepared; the reviewed complete tree has only a README and teaser image, with no weights or inference implementation. T2 is currently provided through the hosted API. Closed-weight status is time-bounded to this evidence, not a claim that a future public release cannot happen.

Classification evidence (opens in a new tab)

Lifecycle dates

Smart Topology image API introduced13 Jul 2026Source (opens in a new tab)
Text-to-3D Smart Topology introduced13 Aug 2026Source (opens in a new tab)

Sources reviewed