Hy4 Flash Base
- In Input modalities: TextImage
- Open weights
Prices
0 offerings · prices as listedNo priced offering is listed for this model yet.
Details
- Licence
- tencent-hy-license
- Family
- hy-flash-base
Run locally
Weights: tencent/Hy4-flash-base- Parameters 219.4B
- Active 12B · 512 experts, 16 per token
- Native precision BF16
- Layers 42
- KV cache per token 47 KB
- Config positions 1024K (config.json max_position_embeddings)
- Needs
- ≈ 114.4 GB (weights 108.6 + KV cache 0.4 + overhead 5.4)
- Usable
- 21.6 GB (90% of its memory)
- Try
- IQ2_XXS runs (needs system memory), but below 4 bits quality drops noticeably
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 408.7 GB | ≈ 429.5 | ✕Doesn’t fit | — |
| Q8_0 | 217.1 GB | ≈ 228.3 | ✕Doesn’t fit | — |
| FP8 | 204.3 GB | ≈ 214.9 | ✕Doesn’t fit | — |
| Q6_K | 167.8 GB | ≈ 176.6 | ✕Doesn’t fit | — |
| Q5_K_M | 145.3 GB | ≈ 153.0 | ✕Doesn’t fit | — |
| Q5_0 | 140.5 GB | ≈ 147.9 | ✕Doesn’t fit | — |
| Q4_K_M | 124.1 GB | ≈ 130.7 | ✕Doesn’t fit | — |
| Q4_0 | 114.9 GB | ≈ 121.1 | ✕Doesn’t fit | — |
| IQ4_XS | 108.6 GB | ≈ 114.4 | ✕Doesn’t fit | — |
| Q3_K_Mquality drops | 99.9 GB | ≈ 105.2 | ✕Doesn’t fit | — |
| IQ3_XXSquality drops | 78.2 GB | ≈ 82.4 | ✕Doesn’t fit | — |
| Q2_Kquality drops | 76.6 GB | ≈ 80.8 | ✕Doesn’t fit | — |
| IQ2_XXSquality drops | 52.6 GB | ≈ 55.6 | ○Needs system memory 24–36 tok/s | — |
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 NVIDIA B200 (HGX B200); unified memory Apple M5 Ultra (64-core GPU) · 256 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/Hy4-flash-base model card (opens in a new tab) huggingface.co · checked Nov 18, 2026
- tencent/Hy4-flash-base config.json (opens in a new tab) huggingface.co · checked Nov 18, 2026
- tencent/Hy4-flash-base files (opens in a new tab) huggingface.co · checked Nov 18, 2026
- tencent/Hy4-flash-base README.md (opens in a new tab) huggingface.co · checked Nov 18, 2026
- tencent/Hy4-flash-base model card (opens in a new tab) huggingface.co · checked Nov 18, 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