Table with columns: Tensors, Format| Tensors | Format |
|---|
| attention q, k, o (60 layers) | bfloat8_b, 8.5 bits a weight |
| attention v (50 layers) | bfloat8_b, 8.5 bits a weight |
| MLP gate, up, down (60 layers) | bfloat8_b, 8.5 bits a weight |
| LM head (tied to the embedding), embedding, norms and every other tensor | bf16, as in the base model |
embed_vision, vision_tower | bf16, unchanged |
8.85 bits a weight on average, 33.94 GB packed. quant4.json lists each projection's format. Quantized with
GPTQ on the bfloat8_b grid, keeping each block's shared exponent as the packer sets it, calibrated on 65,536
tokens of WikiText-2 and Korean Wikipedia.
Only the language model was quantized and evaluated.
Quickstart
from transformers import AutoModelForCausalLM, AutoTokenizer repo = "Xylolabs/gemma-4-31B-it-qat-q4_0-unquantized-Tenstorrent-Tensix-BFP8"tok = AutoTokenizer.from_pretrained(repo)model = AutoModelForCausalLM.from_pretrained(repo, dtype="bfloat16", device_map="auto") messages = [{"role": "user", "content": "What is the capital of France?"}]inputs = tok.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt", return_dict=True).to(model.device)out = model.generate(**inputs, max_new_tokens=64)print(tok.decode(out[0, inputs["input_ids"].shape[1]:], skip_special_tokens=True))
Run on Tensix
Checked with ttnn 0.81.1 on ttsim, Tenstorrent's Blackhole simulator. Pack one matrix and
check that no value changed (TT-NN reads weights as [in, out]):
import jsonimport torchimport ttnnfrom safetensors import safe_open d = "gemma-4-31B-it-qat-q4_0-unquantized-Tenstorrent-Tensix-BFP8" # hf download Xylolabs/gemma-4-31B-it-qat-q4_0-unquantized-Tenstorrent-Tensix-BFP8 --local-dir gemma-4-31B-it-qat-q4_0-unquantized-Tenstorrent-Tensix-BFP8name = "model.language_model.layers.0.mlp.down_proj.weight"fmt = json.load(open(f"{d}/quant4.json"))["layers"]["0"]["mlp.down_proj"]["format"]dtype = {"bfp4": ttnn.bfloat4_b, "bfp8": ttnn.bfloat8_b}[fmt]shard = json.load(open(f"{d}/model.safetensors.index.json"))["weight_map"][name]with safe_open(f"{d}/{shard}", "pt") as f: w = f.get_tensor(name).t().contiguous() device = ttnn.open_device(device_id=0)w_tt = ttnn.as_tensor(w, dtype=dtype, layout=ttnn.TILE_LAYOUT, device=device, memory_config=ttnn.DRAM_MEMORY_CONFIG)assert torch.equal(ttnn.to_torch(w_tt).float(), w.float())ttnn.close_device(device)
python pack_tensix.py gemma-4-31B-it-qat-q4_0-unquantized-Tenstorrent-Tensix-BFP8 checks every matrix this way
(--device also round-trips it through device DRAM).
python verify_grid.py gemma-4-31B-it-qat-q4_0-unquantized-Tenstorrent-Tensix-BFP8 checks the grid with torch alone.
tt-transformers and tt-forge were not run with this checkpoint.
Evaluation
Table with columns: Bits a weight, Size (GB), Mean KL to BF16, Top-1 (%), Answers right, Lost | Bits a weight | Size (GB) | Mean KL to BF16 | Top-1 (%) | Answers right | Lost |
|---|
| BF16 (google/gemma-4-31B-it-qat-q4_0-unquantized) | 16 | 61.39 | 0 | 100 | 33 of 33 | - |
| this model | 8.85 | 33.938 | 0.00174 | 98.67 | 33 of 33 | 0 |
| Q4_0 (llama.cpp) | 4.60 | 17.651 | 0.01920 | 95.33 | 33 of 33 | 0 |
KL and top-1 cover 8,048 positions of English and Korean text (WikiText-2 test, Korean Wikipedia), in the chat
template. Answers are 33 short prompts with automatic checks. "Lost" means right in BF16 but wrong or broken
here; near ties are not counted. Q4_0 was made from the same checkpoint with llama.cpp v0.9.0 (cd959bae) and
no importance matrix; its sha256 is 141db4794deefb1b40be410f297e53ac17a0742b879edbe98c591210c68b47cd. Q4_0
is the reference because Google ships this QAT checkpoint as Q4_0. Our bar: mean KL at most 1.5 times Q4_0's
and at most one more lost answer. Passed: 0.09 times Q4_0's KL, 0 lost against 0.
These numbers measure closeness to BF16, not task quality. Speed is not reported here.
License
Apache 2.0, as the base model's card says; its repository has no LICENSE file, so LICENSE holds the standard
text. The changes are listed in NOTICE. pack_tensix.py and verify_grid.py are by Xylolabs under the same
license. Provided as is, without warranty. Not affiliated with or endorsed by Tenstorrent.
Citation
@misc{gemmateam2026gemma4, title={Gemma 4 Technical Report}, author={Gemma Team}, year={2026}, eprint={2607.02770}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2607.02770},}