whatmodel.sh / image / HunyuanImage 2.1
HunyuanImage 2.1
↗ share
+ contribute
HunyuanImage 2.1 is an open-weights image model from Tencent, released September 2025. Its published size is 17B parameters. Licence: tencent-hunyuan-community.
Model information sources
Release date ↗ · Weights licence ↗ · Parameter count ↗
Base text-to-image diffusion transformer only; excludes the MLLM and character-aware text encoders, the refiner stage and the VAE.
Score within the LMArena text-to-image comparison set only.
$ how it compares
The image models ranked either side of HunyuanImage 2.1, out of 197 with a score, then more from Tencent.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 63 |  Z-Image (Base) | 58±2 | api $0.010/image +local | ||
| 64 |  Wan 2.6 Image | 58±4 | api $0.030/image | ||
| 65 | HunyuanImage 2.1 | 58±2 | api $0.10/image +local | ||
| 66 |  Qwen-Image-Edit-2511 | 58±2 | api $0.030/image +local | ||
| 67 |  Recraft V4.1 Flash | 57±3 | api $0.0070/image | ||
| more from Tencent | |||||
| 38 | HunyuanImage 3.0-Instruct | 62±3 | api $0.090/image +local | ||
| 68 | HunyuanImage 3.0 | 57±2 | api $0.080/image +local | ||
$ what it's good at
text to image▂▅▅▃▄▄█▃▁▂▁▁ #57 of 126 * · 16% of its score · 1 test
Trails LLaDA-Image, Z-Image-Turbo, Qwen-Image-2512 here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| ImageTime (prompt-only, overall) | 49.8% ✎ | ▄·····▄█▄·▄█ beats 1/7 | 16% |
other tests1 not in a task · 81% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| VVRBench accuracy (10,000 verifiable synthetic-scene tasks) | 18.8% ✎ | ▃·█▅▃·▅▃···▃ beats 7/10 | 81% |
$ where to run it
Through an API
api · fal.ai · $0.10/image · fal.ai ↗ · price sourceUSD for one 1-megapixel output: fal bills this endpoint $0.1 per megapixel (price = rate × 1 MP; fal's rounding for 1024×1024 ≈ 1.05 MP is not stated; larger outputs cost proportionally more). Product: fal.ai serverless text-to-image. API model: fal-ai/hunyuan-image/v2.1/text-to-image (Hunyuan Image 2.1). Settings: 1 MP output, one image. Price basis: published per-megapixel list rate × 1 MP. Currency: USD. Not a Tencent direct tariff. fal.ai official model page, observed 2026-09-26.
Other ways to run it
- Run it as published · needs ~64.2 GB, NVIDIA only · from huggingface.co/hunyuanvideo-community
uv run --with diffusers --with transformers --with accelerate --with torch python -c "import torch; from diffusers import HunyuanImagePipeline; p=HunyuanImagePipeline.from_pretrained('hunyuanvideo-community/HunyuanImage-2.1-Diffusers', torch_dtype=torch.bfloat16).to('cuda:0'); p('a cute cartoon penguin plush toy painting the Mona Lisa in a studio', num_inference_steps=50, height=2048, width=2048, generator=torch.Generator(device='cuda:0').manual_seed(649151)).images[0].save('out.png')"
Official weights and documentation: huggingface.co/tencent ↗
$ how the score was worked out · Verifiable Visual Rewards Tra… 81%, Can Image Models Imagine Time… 16%
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 |
|---|---|---|
 Verifiable Visual Rewards Transfer from Synthetic Scenes to Natural Prompts (VVRBench), arXiv:2609.35641v1 | 1 | 81% |
| Can Image Models Imagine Time? ImageTime: A Novel Benchmark for Probing Visual World Modeling Through Spatiotemporal Consistency | 1 | 16% |
| prior · pull toward the middle | 2% | |
| weighted average | = 57.8 |
$ 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