turintech/gemma-4-E4B-it-Artemis-GGUF
Text Generation • 7B • Updated • 1.55k • 8
Compressed LLMs
Third-party LLMs in smaller GGUF builds from TurinTech, built with Artemis. See each card for tests and hardware.
Note Google Gemma 4 E4B chat model. The default 4.2 GB file predicts the same next token as the original 98.4% of the time. Text only. Measured 3.3 GiB GPU plus 4.4 GiB system memory; Unsloth's UD-Q4_K_XL of the same size tests about the same.
turintech/gemma-4-E2B-it-Artemis-GGUF
Text Generation • 5B • Updated • 2.03k • 8Note Google Gemma 4 E2B chat model. The default 2.3 GB file predicts the same next token as the original 95.8% of the time. Text only. Memory not measured yet; if 2.6 GB fits, Unsloth's UD-Q4_K_XL is closer.
turintech/GLM-5.2-MaxMin108-GGUF
Text Generation • 325B • Updated • 288 • 11Note GLM-5.2 753B, expert-pruned: 108 of 256 experts per layer kept. The default -fast file is 95.8 GiB and needs a 128 GB machine (one GB10). MMLU 61.84 against 87.29 for the parent. Text, English only.
-
turintech/Qwen3.5-0.8B-Artemis-GGUF
Text Generation • 0.8B • Updated • 632 • 4 -
turintech/Qwen3.5-2B-Artemis-GGUF
Text Generation • 2B • Updated • 518 • 2 -
turintech/Qwen3.5-4B-Artemis-GGUF
Text Generation • 4B • Updated • 584 • 1 -
turintech/Qwen3.5-9B-Artemis-GGUF
Text Generation • 9B • Updated • 596 • 1