MiniOcDevKit/README at feat/llama.cpp · Light4Code/MiniOcDevKit · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
# MINIOCDEVKIT

## Model Benchmarks

The benchmark task prompt used is "Create a very basic todo cli app with rust. It should handle pending and complete tasks.".

Model provider: LLama.cpp local server on Mac M1 Pro 32GB.

| Model Name                         | Tokens/s | Provider Context Size   | OC Context Size | Success | Notes                                                                                  |
|------------------------------------|----------|-------------------------|-----------------|---------|----------------------------------------------------------------------------------------|
| Qwen3.6-27B-UD-IQ3_XXS.gguf        | 8.5      | 32768                   | 28672           | Yes     | [Qwen3.6-27B-GGUF](https://huggingface.co/unsloth/Qwen3.6-27B-GGUF)                    |
| Qwen3.6-27B-MTP-UD-IQ3_XXS.gguf    | 6.9      | 32768                   | 28672           | Yes     | [Qwen3.6-27B-MTP-GGUF](https://huggingface.co/unsloth/Qwen3.6-27B-MTP-GGUF)            |
| Qwen3.6-27B-MTP-IQ4_XS.gguf        | 6.7      | 32768                   | 28672           | ?       | [Qwen3.6-27B-MTP-GGUF](https://huggingface.co/unsloth/Qwen3.6-27B-MTP-GGUF)            |
| Qwen3.6-35B-A3B-MXFP4_MOE.gguf     | 9.25     | 32768                   | 28672           | Yes     | [Qwen3.6-35B-A3B-GGUF](https://huggingface.co/unsloth/Qwen3.6-35B-A3B-GGUF)            |
| Qwen3.5-4B-MTP-IQ4_XS.gguf         | ?        | 32768                   | 28672           | Yes     | [Qwen3.5-4B-MTP-GGUF](https://huggingface.co/unsloth/Qwen3.5-4B-MTP-GGUF)              |
| Qwen3.6-35B-A3B-MTP-MXFP4_MOE.gguf | 9.95     | 32768                   | 28672           | Yes     | [Qwen3.6-35B-A3B-MTP-GGUF](https://huggingface.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF)    |