WeMM Video 9B NVFP4
- Open weights
Prices
0 offerings · prices as listedNo priced offering is listed for this model yet.
Details
- Family
- wemm-video-nvfp
Run locally
Weights: tencent/WeMM-Video-9B-NVFP4- Parameters 10.5B
- Dense
- Native precision BF16
- Layers 32
- KV cache per token 32 KB
- Config positions 256K (config.json max_position_embeddings)
- Needs
- ≈ 11.1 GB (weights 10.4 + KV cache 0.3 + overhead 0.5)
- Usable
- 21.6 GB (90% of its memory)
- Generation speed
- 44–66 tok/s
- Longest context
- 256K (the model’s limit)
- First token
- about 0.34–0.51 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native) | 19.5 GB | ≈ 20.8 | ●Runs well 24–36 tok/s | 34K |
| Q8_0 | 10.4 GB | ≈ 11.1 | ●Runs well 44–66 tok/s | 256K |
| FP8 | 9.8 GB | ≈ 10.5 | ●Runs well 47–70 tok/s | 256K |
| Q6_K | 8.0 GB | ≈ 8.8 | ●Runs well 57–85 tok/s | 256K |
| Q5_K_M | 6.9 GB | ≈ 7.7 | ●Runs well 65–98 tok/s | 256K |
| Q5_0 | 6.7 GB | ≈ 7.5 | ●Runs well 67–101 tok/s | 256K |
| Q4_K_M | 5.9 GB | ≈ 6.7 | ●Runs well 76–114 tok/s | 256K |
| Q4_0 | 5.5 GB | ≈ 6.2 | ●Runs well 82–123 tok/s | 256K |
| IQ4_XS | 5.2 GB | ≈ 5.9 | ●Runs well 86–129 tok/s | 256K |
| Q3_K_Mquality drops | 4.8 GB | ≈ 5.5 | ●Runs well 93–140 tok/s | 256K |
| IQ3_XXSquality drops | 3.7 GB | ≈ 4.5 | ●Runs well 118–177 tok/s | 256K |
| Q2_Kquality drops | 3.7 GB | ≈ 4.4 | ●Runs well 120–180 tok/s | 256K |
| IQ2_XXSquality drops | 2.5 GB | ≈ 3.3 | ●Runs well 170–255 tok/s | 256K |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 3080; unified memory Apple M5 · 16 GB.
Run commands
The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-21. How it’s estimated
Cited scores
From third-party leaderboards, not measured by usThis model isn’t on any leaderboard we cite yet.
Changes
All changes →- New language model
Sources
- tencent/WeMM-Video-9B-NVFP4 model card (opens in a new tab) huggingface.co · checked Nov 20, 2026
- tencent/WeMM-Video-9B-NVFP4 config.json (opens in a new tab) huggingface.co · checked Nov 20, 2026
Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data
Tencent · China · tencent.com
Related models
Snapshots, the same family and other models from this provider