Agnes Video H3 MoE vs AURA 1.3B
Native audio: Agnes Video H3 MoE. Not enough arena votes yet for a quality score, so this is a spec comparison.
Agnes AI
Agnes Video H3 MoE
- Arena—
- From—
- Max length—
Kyutai
AURA 1.3B
- Arena—
- From—
- Max length—
Add a model
Make it a three-way comparison.
Where they differ
Quality is the arena score from blind, side-by-side human votes. Price, length and resolution come from each lab’s own listings.
- NewestSame release dateAgnes Video H3 MoE Nov 20, 2026 · AURA 1.3B Nov 20, 2026
- Widest inputsSame inputsAgnes Video H3 MoE: Text prompts, Reference images · AURA 1.3B: Reference images, Audio
- Native audioAgnes Video H3 MoEGenerates an audio track with the video: Agnes Video H3 MoE yes · AURA 1.3B unconfirmed
- Image to videoBoth animate imagesStarts from an uploaded image: Agnes Video H3 MoE yes · AURA 1.3B yes
- Self-hostingAll open weightsDownloadable weights (MiniMax H3 Community License and CC BY-NC 4.0)
- Most API hostsSame reachAgnes Video H3 MoE: 0 · AURA 1.3B: 0
Every spec in one table
Dashes mean the data is not published.
| Specification | ||
|---|---|---|
| Task | Audiovisual generation | Speech-driven portrait animation |
| Published capabilities | Experimental MiniMax H3 FL2VA adaptation that replaces selected feed-forward layers with routed sparse experts. | A four-step causal model animates a portrait from streaming speech, rendering chunks from the available audio. |
| Model size | Not published | 1.3B parameters |
| Access | Public composite package with original weights, learned router and adapter, and reference loader | Public student checkpoints and inference code |
| Result | Video with synchronized audio | Not published |
| Operating limits | Supports an initial frame, an ending frame or both with the publisher’s reference loader. | Kyutai reports 20.7 fps end to end on one H100 at 576p for 16 Hz playback. |
| Sources | Model weights | Model weights |
| Quality (blind human votes) | ||
| Documented capabilities | Experimental MiniMax H3 FL2VA adaptation that replaces selected feed-forward layers with routed sparse experts. | A four-step causal model animates a portrait from streaming speech, rendering chunks from the available audio. |
| Model parameters | Not published | 1.3B parameters |
| Text to video arenaElo-style score · higher is better | Not published | Not published |
| Image to video arenaElo-style score · higher is better | Not published | Not published |
| Price | ||
| Starting price | Not published | Not published |
| Top price | Not published | Not published |
| 10-second clip | Not published | Not published |
| Extra charges | Not published | Not published |
| How it is billed | Not published | Not published |
| Output | ||
| Produces | Video with synchronized audio | Video |
| Resolutions | Not published | 576p |
| Highest resolution | Not published | |
| Clip length | Not published | Not published |
| Longest clip | Not published | Not published |
| Sound | Always generated | Not published |
| Frame control | First and last frame | Not published |
| Aspect ratios | Not published | Not published |
| Inputs | ||
| Text prompts | Yes | No |
| Start image (image to video) | Yes | Yes |
| Video input | No | No |
| Audio input | No | Yes |
| References | Not published | Not published |
| Access | ||
| Weights | OpenMiniMax H3 Community License | OpenCC BY-NC 4.0 |
| API model ID | Not published | Not published |
| API hosts | Not published | Not published |
| Released | Nov 20, 2026 | Nov 20, 2026 |
| Documentation | Model weights | Model weights |
| Specification sources | Not published | Not published |
| Specs checked | Not published | Not published |
Which should you choose?
Which is better, Agnes Video H3 MoE or AURA 1.3B?
Neither model has enough arena votes for a reliable quality score yet, so this page compares specs only. Run the same three or four prompts through each and judge the results.
What is the difference between Agnes Video H3 MoE and AURA 1.3B?
Agnes Video H3 MoE is a video model from Agnes AI (Nov 20, 2026), taking text prompts and reference images and producing video with synchronized audio. AURA 1.3B is a video model from Kyutai (Nov 20, 2026), taking reference images and audio and producing video.
Which is newer?
Agnes Video H3 MoE Nov 20, 2026; AURA 1.3B Nov 20, 2026.
Which generates video with sound?
Generates an audio track with the video: Agnes Video H3 MoE yes; AURA 1.3B unconfirmed.
Can I run any of these on my own hardware?
All open weights. Downloadable weights (MiniMax H3 Community License and CC BY-NC 4.0).