Anthropic
OpenAI
Google
Meta
xAI
MistralWhich inputs and outputs matter?
Distinguish image understanding, image generation, OCR, visual question answering, and audio. The domain labels are not interchangeable.
Purpose guide · Multimodal
Which AI models belong on a multimodal shortlist?
Catalog current through July 14, 2026
12 models
Source-backed models whose Epoch records explicitly include a multimodal, vision, image, or audio domain. This is a capability map, not a quality ranking.
Selection rule
Included when the Epoch record explicitly lists multimodal, vision, image generation, or audio as a domain. Domain membership does not prove support for every media type or task.
Task-fit shortlist
Evidence first. Your evaluation next.
Ordered by release date for readability, newest first. The order is not a score or recommendation.
- 01
- Matching evidence
- Multimodal
- Domain
- Language · Multimodal
- Access
- API access; closed weights
- 02
- Matching evidence
- Multimodal · Vision
- Domain
- Multimodal · Language · Vision
- Access
- API access; closed weights
- 03
- Matching evidence
- Image generation
- Domain
- Image generation
- Access
- API access; closed weights
- 04
- Matching evidence
- Audio
- Domain
- Audio
- Access
- API access; closed weights
- 05
- Matching evidence
- Multimodal · Vision
- Domain
- Multimodal · Language · Vision
- Access
- API access; closed weights
- 06
- Matching evidence
- Vision
- Domain
- Language · Vision
- Access
- API access; closed weights
- 07
- Matching evidence
- Image generation
- Domain
- Image generation
- Access
- API access; closed weights
- 08
- Matching evidence
- Multimodal · Vision
- Domain
- Language · Multimodal · Vision
- Access
- API access; closed weights
- 09
- Matching evidence
- Multimodal · Vision
- Domain
- Multimodal · Language · Vision
- Access
- Open weights (restricted use); open weights
- 10
- Matching evidence
- Multimodal · Vision
- Domain
- Multimodal · Language · Vision
- Access
- Open weights (restricted use); open weights
- 11
- Matching evidence
- Multimodal · Vision
- Domain
- Multimodal · Vision · Language
- Access
- API access; closed weights
- 12
- Matching evidence
- Multimodal · Vision
- Domain
- Multimodal · Language · Vision
- Access
- Open weights (non-commercial); open weights
Decision criteria
Questions the source can help you ask.
These checks narrow an evaluation plan. They do not replace hands-on testing.
Is the workflow mixed-media?
Test the full handoff between text, files, images, and tools. A multimodal label alone does not establish format support or reliable cross-modal reasoning.
What must you test yourself?
Evaluate media quality, instruction following, file limits, safety, latency, and cost on representative assets. Those measurements are not in this catalog.
Evidence boundary
What this guide does not know.
No task-level quality, price, context window, latency, or reliability measurement is present in the Epoch catalog. We therefore publish no ranking or “best” label.
Catalog source: Epoch AI Data on AI Models, retrieved 2026-07-14, CC-BY. Artifact sha256:3c663b17ec989f00ef5f994214b878a914118e2d1fb6077e23329a318209eb86.
Each listed model links to its model-level source above. See the publication source ledger and methodology for the wider evidence policy.