# Gemma4-31B Uncensored: 1M Context + MTP + Vision
[HauhauCS/Gemma4-31B-QAT-Uncensored-HauhauCS-Balanced-MTP](https://huggingface.co/HauhauCS/Gemma4-31B-QAT-Uncensored-HauhauCS-Balanced-MTP) (31B dense, Google QAT checkpoint) with a **1,048,576-token context baked in** (4x the native 262,144), shipping with its MTP speculative-decoding draft head and vision tower. All numbers below were measured on these exact files.
| Capability | Status |
|---|---|
| 1M context | 10/10 through 131K; 262K to 1M rungs in progress, card will update |
| MTP speculative decoding | 69.2 to 101.0 tok/s (+46%), acceptance 0.658 (measured on this trunk, RTX 5090) |
| Vision | Verified July 6, 2026: reads image text and identifies objects |
| Uncensored | HauhauCS Balanced abliteration; trunk weights bit-identical to the source release |
Perfect scores as far as a 32 GB card could take a dense 31B at f16 KV (131K). The 262K through 1M rungs are running on a 128 GB Mac at publish time and this card will be updated as each lands. Side note: DeepReinforce's Ornith-1.0 family description lists an unreleased 31B Dense variant built on this same Gemma 4 trunk; only their Qwen-based models have shipped.
## MTP speculative decoding
The draft head predicts ahead and the trunk verifies every token, so output is identical to standard decoding, only faster. Measured speedup on this uncensored trunk beats the ~35 percent claimed upstream.
## Files
| File | Size | Role |
|---|---|---|
| `gemma4-31b-uncensored-1M-Q4.gguf` | 18.7 GB | Trunk, 1M baked, QAT 4-bit |
| `mtp-gemma-31b.gguf` | 280 MB | MTP draft head, pair with `-md` |
| `mmproj-gemma31b-hauhau.gguf` | 1.2 GB | Vision tower, pair with `--mmproj` |
| `niah_heatmap.png`, `mtp_speedup.png`, `results.jsonl` | small | Verification evidence |
## Every file, every mirror
Nothing was discontinued: every quant is one click away. Hugging Face carries the curated picks, ModelScope always carries everything, and Ollama serves ready-to-run tags.
| File | Size | Hugging Face | ModelScope | Ollama |
|---|---|---|---|---|
| `gemma4-31b-uncensored-1M-Q4.gguf` | 18.7 GB | [download](https://huggingface.co/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF/resolve/main/gemma4-31b-uncensored-1M-Q4.gguf) | [download](https://www.modelscope.ai/models/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF/resolve/master/gemma4-31b-uncensored-1M-Q4.gguf) | - |
| `mmproj-gemma31b-hauhau.gguf` | 1.2 GB | [download](https://huggingface.co/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF/resolve/main/mmproj-gemma31b-hauhau.gguf) | [download](https://www.modelscope.ai/models/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF/resolve/master/mmproj-gemma31b-hauhau.gguf) | - |
| `mtp-gemma-31b.gguf` | 280 MB | [download](https://huggingface.co/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF/resolve/main/mtp-gemma-31b.gguf) | [download](https://www.modelscope.ai/models/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF/resolve/master/mtp-gemma-31b.gguf) | - |
## Run it
llama.cpp, everything on:
```bash
llama-server -m gemma4-31b-uncensored-1M-Q4.gguf \
-c 1048576 -np 1 --jinja \
-md mtp-gemma-31b.gguf --spec-type draft-mtp --spec-draft-n-max 3 \
--mmproj mmproj-gemma31b-hauhau.gguf
```
Ollama (1M and vision work; Ollama has no speculative decoding yet, so the MTP head adds no speed there):
```
FROM ./gemma4-31b-uncensored-1M-Q4.gguf
RENDERER gemma4
PARSER gemma4
PARAMETER num_ctx 262144
```
The RENDERER and PARSER lines avoid imported-GGUF template bugs under tool-heavy use. Raise `num_ctx` as memory allows.
## How this was built
YaRN rope-scaling metadata (factor 4.0 over native 262,144) baked into the GGUF header with gguf-py; weights are bit-identical to the HauhauCS release, no fine-tuning. Gemma 4's dual-rope design takes YaRN on its global-attention layers. Certification harness: 10 needles per rung at depths 5 to 95 percent, temperature 0, seeded prompts, f16 KV only. Method and tooling: [github.com/satindergrewal/aviary-1m](https://github.com/satindergrewal/aviary-1m).
For base capability benchmarks see Google's official Gemma 4 cards; uncensoring quality versus the official trunk has not been independently benchmarked here.
## Credits
Base model and QAT: Google (Gemma license; its terms flow down to these files). Uncensoring and packaging: [HauhauCS](https://huggingface.co/HauhauCS/Gemma4-31B-QAT-Uncensored-HauhauCS-Balanced-MTP). MTP head: Unsloth (via the HauhauCS repo). 1M YaRN extension, benchmarking, and certification: [SatGeze](https://huggingface.co/satgeze).
Sister repos: [12B](https://huggingface.co/satgeze/Gemma4-12B-Uncensored-HauhauCS-1M-GGUF) | [26B-A4B](https://huggingface.co/satgeze/Gemma4-26B-A4B-Uncensored-HauhauCS-1M-GGUF) | [31B](https://huggingface.co/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF) | [Qwen3.6-35B](https://huggingface.co/satgeze/Qwen3.6-35B-Uncensored-HauhauCS-1M-GGUF)
Mirrors: Hugging Face | [ModelScope](https://www.modelscope.ai/models/satgeze/Gemma4-31B-Uncensored-HauhauCS-1M-GGUF)