---
license: apache-2.0
arxiv: 2608.23283
language:
  - en
  - zh
base_model: apodex/Apodex-1.1-mini
base_model_relation: quantized
pipeline_tag: text-generation
library_name: gguf
tags:
  - gguf
  - llama.cpp
  - qwen3.5
  - moe
  - multimodal
  - mtp
  - apodex
---

# Apodex-1.1-mini GGUF

GGUF quantizations of [apodex/Apodex-1.1-mini](https://huggingface.co/apodex/Apodex-1.1-mini), a 35.95B-parameter Qwen3.5 MoE model for research, data, files, code, and tool-driven work.

All model files retain the checkpoint's native MTP/NextN head. For image input, use the included F16 vision projector.

## Upstream benchmarks

![Apodex-1.1 benchmark results](assets/apodex-1.1-benchmarks.png)

*Results reported by Apodex AI from evaluations of the original Apodex-1.1 model using FrontierAgent, its agent harness and evaluation suite.*

![Apodex-1.1-mini benchmark results](assets/apodex-1.1-mini-benchmarks.png)

*Results reported by Apodex AI from evaluations of the original Apodex-1.1-mini checkpoint using FrontierAgent. With the Agent Team workflow, it scored 27.7 on APEX-Agent, 50.2 on FrontierFinance, and 51.7 on FrontierScience-Research.*

## GGUF files

| Quantization | File | Size | Notes |
| --- | --- | ---: | --- |
| IQ1_M | [Apodex-1.1-mini-IQ1_M.gguf](Apodex-1.1-mini-IQ1_M.gguf) | 8.82 GB | Ultra-low-memory option |
| IQ2_M | [Apodex-1.1-mini-IQ2_M.gguf](Apodex-1.1-mini-IQ2_M.gguf) | 12.2 GB | Smallest recommended option |
| Q2_K | [Apodex-1.1-mini-Q2_K.gguf](Apodex-1.1-mini-Q2_K.gguf) | 13.2 GB | Maximum-compression K-quant |
| IQ3_M | [Apodex-1.1-mini-IQ3_M.gguf](Apodex-1.1-mini-IQ3_M.gguf) | 16.0 GB | Importance-aware compact option |
| Q3_K_M | [Apodex-1.1-mini-Q3_K_M.gguf](Apodex-1.1-mini-Q3_K_M.gguf) | 17.2 GB | Standard Q3 K-quant |
| Q4 Selective | [Apodex-1.1-mini-Q4_Selective-Q4_K_M.gguf](Apodex-1.1-mini-Q4_Selective-Q4_K_M.gguf) | 19.8 GB | Compact custom Q4 alternative |
| Q4_K_S | [Apodex-1.1-mini-Q4_K_S.gguf](Apodex-1.1-mini-Q4_K_S.gguf) | 20.4 GB | Smaller Q4 option |
| Q4_K_M | [Apodex-1.1-mini-Q4_K_M.gguf](Apodex-1.1-mini-Q4_K_M.gguf) | 21.7 GB | Recommended default |
| Q5_K_S | [Apodex-1.1-mini-Q5_K_S.gguf](Apodex-1.1-mini-Q5_K_S.gguf) | 24.6 GB | Smaller Q5 option |
| Q5_K_M | [Apodex-1.1-mini-Q5_K_M.gguf](Apodex-1.1-mini-Q5_K_M.gguf) | 25.3 GB | Higher-quality Q5 option |
| Q6_K | [Apodex-1.1-mini-Q6_K.gguf](Apodex-1.1-mini-Q6_K.gguf) | 29.2 GB | High-quality option |
| Q8_0 | [Apodex-1.1-mini-Q8_0.gguf](Apodex-1.1-mini-Q8_0.gguf) | 37.8 GB | Near-lossless reference quantization |
| Vision projector | [mmproj-Apodex-1.1-mini-F16.gguf](mmproj-Apodex-1.1-mini-F16.gguf) | 899 MB | Required for image input |

The IQ builds use the included [importance matrix](imatrix/Apodex-1.1-mini-wiki-100x512.imatrix). Q4 Selective is a compact custom Q4 alternative that selectively retains higher precision in key weights.

## Usage

Use a current [llama.cpp](https://github.com/ggml-org/llama.cpp) build with the included [chat template](chat_template.jinja).

```bash
llama-cli \
  -m Apodex-1.1-mini-Q4_K_M.gguf \
  -c 4096 -n 512 --temp 1.0 --top-p 0.95 \
  --jinja --chat-template-file chat_template.jinja \
  -p "Explain why reproducible builds matter."
```

For MTP speculative decoding:

```bash
llama-server \
  -m Apodex-1.1-mini-Q4_K_M.gguf \
  --spec-type draft-mtp --spec-draft-n-max 3 \
  --jinja --chat-template-file chat_template.jinja
```

For image input:

```bash
llama-mtmd-cli \
  -m Apodex-1.1-mini-Q4_K_M.gguf \
  -mm mmproj-Apodex-1.1-mini-F16.gguf \
  --image image.jpg -p "Describe this image."
```

## Source

- Model: [apodex/Apodex-1.1-mini](https://huggingface.co/apodex/Apodex-1.1-mini)
- Chat template revision: [`62583b4`](https://huggingface.co/apodex/Apodex-1.1-mini/commit/62583b464707ad58848bc2ab066747d8d5b02533)
- License: [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)
- Checksums: [SHA256SUMS](SHA256SUMS)