Benchmarks
How open-weight models stack up — across reasoning, code, vision, and image generation. Cached from public sources and refreshed every 12 hours.
Artificial Analysis Intelligence Index — a blended measure of reasoning, knowledge, and coding across open-weight and proprietary language models.
Best value · intelligence vs. price
How much capability each open model delivers per dollar. Models on the frontier (top-left) lead on value.
Every board
Artificial Analysis Intelligence Index — a blended measure of reasoning, knowledge, and coding across open-weight and proprietary language models.
| # | Model | Intelligence |
|---|---|---|
| 1 | Kimi K3 Moonshot AI | 57.1 |
| 2 | GLM-5.2 Zhipu AI | 51.1 |
| 3 | DeepSeek V4 Flash 0731 DeepSeek | 49.9 |
| 4 | DeepSeek V4 Pro DeepSeek | 44.3 |
| 5 | Kimi K2.6 Moonshot AI | 44.2 |
Human-preference Elo for text-to-image models from the Artificial Analysis Arena — open-weight and proprietary, ranked head-to-head.
| # | Model | Elo |
|---|---|---|
| 1 | HiDream-O1-Image-Dev-2604 HiDream | 1,188 |
| 2 | ERNIE Image Baidu | 1,166 |
| 3 | ERNIE Image Turbo Baidu | 1,163 |
| 4 | FLUX.2 [dev] Black Forest Labs | 1,154 |
| 5 | FLUX.2 [dev] Turbo Fal | 1,154 |
Human-preference Elo for instruction-based image editing models, ranked head-to-head in the Artificial Analysis Arena.
| # | Model | Elo |
|---|---|---|
| 1 | HunyuanImage 3.0 Instruct Tencent | 1,223 |
| 2 | HiDream-O1-Image HiDream | 1,202 |
| 3 | FLUX.2 [klein] 9B Black Forest Labs | 1,165 |
| 4 | FLUX.2 [dev] Turbo Fal | 1,151 |
| 5 | HiDream-O1-Image-Dev HiDream | 1,139 |
Human-preference Elo for text-to-video models from the Artificial Analysis Arena — the fastest-moving open-vs-proprietary race in generative AI.
| # | Model | Elo |
|---|---|---|
| 1 | LTX-2.3 Fast Lightricks | 1,123 |
| 2 | LTX-2 Fast Lightricks | 1,120 |
| 3 | Wan 2.2 A14B Qwen · Alibaba | 1,109 |
| 4 | HunyuanVideo-1.5 Tencent | 1,019 |
| 5 | Wan 2.1 14B Qwen · Alibaba | 1,013 |
Human-preference Elo for image-to-video models, ranked head-to-head in the Artificial Analysis Arena.
| # | Model | Elo |
|---|---|---|
| 1 | LTX-2 Fast Lightricks | 1,181 |
| 2 | LTX-2.3 Fast Lightricks | 1,159 |
| 3 | HunyuanVideo-1.5 Tencent | 1,125 |
| 4 | Wan 2.2 A14B Qwen · Alibaba | 1,107 |
| 5 | LTX Video v0.9.7 13B Lightricks | 1,036 |
Human-preference Elo for text-to-speech models from the Artificial Analysis Arena — open-weight and proprietary voices, ranked head-to-head.
| # | Model | Elo |
|---|---|---|
| 1 | Kokoro 82M v1.0 Kokoro | 1,055 |
| 2 | Maya1 Maya Research | 1,041 |
| 3 | Higgs Audio V3 TTS Boson AI | 1,040 |
| 4 | Chatterbox Resemble AI | 1,012 |
| 5 | Zonos-v0.1 Zyphra | 1,000 |
Aggregate of contamination-resistant benchmarks — IFEval, BBH, MATH, GPQA, MUSR, MMLU-Pro — for open-weight language models. The final snapshot of Hugging Face's Open LLM Leaderboard.
| # | Model | Avg. |
|---|---|---|
| 1 | MaziyarPanahi/calme-3.2-instruct-78b | 52.1 |
| 2 | MaziyarPanahi/calme-3.1-instruct-78b | 51.3 |
| 3 | dfurman/CalmeRys-78B-Orpo-v0.1 | 51.2 |
| 4 | MaziyarPanahi/calme-2.4-rys-78b | 50.8 |
| 5 | huihui-ai/Qwen2.5-72B-Instruct-abliterated | 48.1 |
The most-downloaded open-weight models on Hugging Face over the last 30 days — the community's working set, straight from the Hub and refreshed every 12 hours.
| # | Model | Downloads |
|---|---|---|
| 1 | sentence-transformers/all-MiniLM-L6-v2 | 257.6M |
| 2 | google-bert/bert-base-uncased | 114.9M |
| 3 | cross-encoder/ms-marco-MiniLM-L6-v2 | 89.1M |
| 4 | BAAI/bge-small-en-v1.5 BAAI | 72.5M |
| 5 | sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 | 60M |
The most-liked open-weight models on Hugging Face — a durable signal of community esteem spanning every modality, refreshed every 12 hours.
| # | Model | Likes |
|---|---|---|
| 1 | Black Forest Labs | 13.9K |
| 2 | deepseek-ai/DeepSeek-R1 DeepSeek | 13.5K |
| 3 | Moonshot AI | 9.8K |
| 4 | stabilityai/stable-diffusion-xl-base-1.0 Stability AI | 8K |
| 5 | CompVis/stable-diffusion-v1-4 | 7K |