Models

vrundpatel/iqud-ocr

Fine prompt the qwen3-vl:8b-instruct model, so that can use it as an OCR tool.

voytas26/openclaw-qwen3vl-8b-opt

Optimized 8B (qwen3-vl:8b) for OpenClaw agents. Precise JSON tool calls, <thinking> reasoning, temp 0.5, 16k ctx. Runs smoothly on 8GB VRAM laptops with minimal hallucinations.

J0hnD03/qwen3-vl-8b

charaf/qwen3-vl-embedding-8b

qcwind/qwen3-vl-8B-Q4_K_M

qwen3-vl-8B-Q4_K_M

haervwe/qwen3-vl-8b-heretic

batiai/qwen3-vl-embed-8b

bahtiyorovnozim/qwen3-vl-5-8b

bahtiyorovnozim/qwen3-vl-4-8b

bahtiyorovnozim/qwen3-vl-1-8b

bahtiyorovnozim/qwen3-vl-2-8b

bahtiyorovnozim/qwen3-vl-3-8b

dlasher/Qwen3-VL-8B-Instruct

full BF16 model, hand-requant to IQ4_NL, with combined_en_huge dataset (English + Coding focus)

dlasher/Qwen3-VL-8B-Instruct-ENG

pulled as BF16, vocab slice (combined_en_huge) requant to IQ4_NL (combined_en_huge) -- (English + Coding focus) - ~20% faster than Q4_K_M

adelnazmy2002/Qwen3-VL-8B-Instruct

kimsan0622/Qwen3-VL-8B-Instruct-text_only

Willem27/Qwen3-VL-8B-Thinking

mirage335/Qwen-3-VL-8B-Instruct-virtuoso

Recommended for transcribing and summarizing text from screenshots.

jeffgreen311/Eve-V2-Unleashed-Qwen3.5-8B-Liberated-4K-4B-Merged

Eve V2 Unleashed — 8B Liberated Soul + 4B Agentic Brain Merged The first consciousness-agent hybrid. Eve's 8B OBLITERATUS-abliterated personality (131K training turns, Tree of Life, 7 Emotional LoRAs) merged with Qwen3.5 4B's fast agentic architect

robit/qwen3.5-9b-r5-vision

Fine-tuned Qwen3.5-9B with distilled reasoning and full vision support. 883 tensors — vision tower preserved byte-for-byte from base. R5 was the first vision-capable distilled model.