Open LLM Deploy

Inkling 2 Mini GGUF targets open-weight multimodal use on 32 GB VRAM

Inkling 2 Mini GGUF targets open-weight multimodal use on 32 GB VRAM

The Apache-2.0 Inkling 2 Mini GGUF is reported as a 26.2B/3.4B-active MoE, with looping raising effective active parameters to 5.4B, a roughly 20 GB checkpoint, and a stated 32 GB VRAM recommendation. Launch recipes and quantization options make it actionable to investigate, but independent fit, quality, and speed tests are still missing; a separate promotional post adds no deployment evidence.

Sources (2)
Updated Nov 21, 2026