whatmodel.sh / video / Wan 2.1 T2V 1.3B
Wan 2.1 T2V 1.3B
↗ share
+ contribute
Wan 2.1 T2V 1.3B is an open-weights video model from Alibaba, released February 2025. Its published size is 1.3B parameters. Licence: apache-2.0.
Model information sources
Release date ↗ · Weights licence ↗ · Parameter count ↗
1.3B = diffusion transformer size per the card's model-configuration table (Model 1.3B, dim 1536, 30 layers); T5 encoder and VAE separate.
$ how it compares
The video models ranked either side of Wan 2.1 T2V 1.3B, out of 139 with a score, then more from Alibaba.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 97 |  Luma Dream Machine 1.0 | 42±7 | comparison only | ||
| 98 |  Sora (original, 2024) | 42±2 | comparison only | ||
| 99 | Wan 2.1 T2V 1.3B | 40±4 | api $0.040/sec +local | ||
| 100 |  Mochi 1 | 40±2 | api price unknown +local | ||
| 101 |  Runway Gen-3 Alpha | 40±3 | comparison only | ||
| more from Alibaba | |||||
| 2 | Wan 3.0 Prime | 80±7 | api $0.14/sec | ||
| 4 | Wan 3.0 | 79±2 | api $0.10/sec | ||
| 9 | HappyHorse 1.1 | 76±3 | api $0.14/sec | ||
$ what it's good at
text to video▆▄▃▄▆▅█▃▂▂▂▄ #46 of 68 · 94% of its score · 2 tests
Trails T2V-Turbo-v2 here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| VBench (T2V total) +1 | 83.3% ✎ | █···▄▄▄····▄ beats 4/5 | 59% |
| WorldJen human preference (Bradley-Terry) | 1355.0 ✎ | █···██·█··██ beats 0/5 | 35% |
$ where to run it
Through an API
api · Replicate · $0.040/sec · replicate.com ↗ · price sourceUSD per second of 480p video, derived from Replicate's price of $0.20 per output video for wan-video/wan-2.1-1.3b, whose page says it generates 5s 480p videos (÷ 5 = $0.04). Product: Replicate model API. API model: wan-video/wan-2.1-1.3b. Settings: resolution=480p (only option), frame_num=81 (default; 16 fps playback). Price basis: published per-output-video rate ÷ the 5-second duration Replicate states; shorter frame_num settings cost the same per video. Currency: USD. Not an Alibaba direct tariff. Replicate official model page, observed 2026-09-26.
Other ways to run it
version
- Run it as published · needs ~8.6 GB · from github.com/leejet
sd-cli -M vid_gen --diffusion-model models/wan2.1_t2v_1.3B_fp16.safetensors --vae models/wan_2.1_vae.safetensors --t5xxl models/umt5-xxl-encoder-Q4_K_M.gguf -p "a lovely cat" --cfg-scale 6.0 --sampling-method euler -v -n "色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走" -W 832 -H 480 --diffusion-fa --video-frames 33 --flow-shift 3.0
Weights and documentation: huggingface.co/Wan-AI ↗
$ how the score was worked out · WorldJen: An End-to-End Multi… 35%, Alibaba Wan technical report… 31%, Training-Free Efficient Video… 28%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 WorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video Models | 1 | 35% |
| Alibaba Wan technical report (arXiv:2503.20314v2), VBench table (own models) · lab-reported | 1 | 31% |
| Training-Free Efficient Video Generation via Dynamic Token Carving | 1 | 28% |
| prior · pull toward the middle | 6% | |
| weighted average | = 40.0 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json