Wan2.2 T2V-A14B
Wan-AI · Wan-AI/Wan2.2-T2V-A14B
Weights license: Apache 2.0 · code and components may use other terms
Quality
1,132 ± 15 · LMArena · text-to-video · Confidence interval 1,116–1,147 · 10,458 votes · Arena writes the name as wan-v2.2-a14b; Wan publishes these weights as Wan2.2-T2V-A14B
Too few of our models are on this leaderboard for a ranking.
Source: LMArena 2026-09-22 · collected 2026-10-25 · dataset · CC BY 4.0
Reported by the publisher
Quality × memory
No comparable quality × memory pair
Image, video and audio models have no shared quality scale, and their memory comes from the publisher's documented minimum path, so quality is not charted against memory.
Runs on
Enough memory 9
A publisher documents a path within this memory, on this platform (CUDA or Apple MPS). Conditional paths need offload or quantization.
Show all 9
One memory tier short 2
The same machines with the next memory size up would have enough.
Hardware profiles grouped by the publisher's documented minimum memory path and platform. Not a measured run.
Consider Wan2.2 T2V-A14B from Wan-AI for the text-to-video task (27B parameters). The official 1280×720 single-GPU command requires at least 80 GB VRAM with model offload and dtype conversion enabled.
This applies to the published --offload_model True --convert_model_dtype command, not a universal minimum. Resolution, frames, precision and runtime change requirements.
Measured speed
No published measurement for this checkpoint
Generation time per clip is shown only when a source publishes it with hardware, build and setup. It is never estimated.
How to run
The official 1280×720 single-GPU command requires at least 80 GB VRAM with model offload and dtype conversion enabled. This applies to the published --offload_model True --convert_model_dtype command, not a universal minimum. Resolution, frames, precision and runtime change requirements. Official hardware evidence
Documented ways to run it
Copied as printed in the linked source (read 2026-10-18); not run by this site.
python generate.py --task t2v-A14B --size 1280*720 --ckpt_dir ./Wan2.2-T2V-A14B --offload_model True --convert_model_dtype --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."torchrun --nproc_per_node=8 generate.py --task t2v-A14B --size 1280*720 --ckpt_dir ./Wan2.2-T2V-A14B --dit_fsdp --t5_fsdp --ulysses_size 8 --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."Evidence level: Publisher documentation linked; no execution reproduced by this site.
Before downloading
- Check the exact checkpoint and license linked here.
- Confirm a compatible runtime and precision or quantization for your platform.
- Allow for context, batch size and any extra media components beyond the weights.
Technical details
What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.
- Checkpoint / version
- Wan-AI/Wan2.2-T2V-A14B
- Task and modality
- text-to-video · video
- Count and basis
- 27B · manufacturer model card
- Released
- Wan-Video/Wan2.2 README (GitHub), Latest News · read 2026-10-09source
- Published repository size
- 126.2 GBNot runtime memory or a minimum download.files
- Published components
- High-noise 14B transformer · Low-noise 14B transformer · Text encoder · VAE
- Runtime and precision
- No local execution verified.
- Commercial use
- Allowed Apache License 2.0 text (SPDX) · read 2026-10-18license text
Every stored field
- Revision checked
- c8c270b13ee05bfa474194ac9fb07a5868a97cea
- Library
- wan2.2
- Hub task tag
- text-to-video
- Formats
- safetensors
- Repository created
- 2025-07-24creation is not the model's release
- Repository updated
- 2025-08-07a commit is not a release date
- Hub downloads (30 days) · likes
- 3,939 · 591this repository only · read 2026-10-25
- Read from the Hub page
- task, task tags, file formatsHub overview and file listing · read 2026-10-18
- Parameters stated by the publisher
- 27,000,000,000model card · “Each expert model has about 14B parameters, resulting in a total of 27B parameters but only 14B active parameters per step” · source · read 2026-10-19
- Catalog state
- No config or weights found on the Hub
Sources
Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.
Catalog reviewed: 2026-09-22.








