AX Engine 7.5.x ships dense 27B past the DRAM ceiling.See the numbers

AX ENGINE / Models

Supported models.

AX Engine supports Qwen 3.5 / 3.6 / 3.8, Gemma 4, GLM, Nemotron, Whisper, and embedding models through a direct-first runtime contract. The qualification pack is Qwen 3.8 27B AXQ 6-bit MTP; other families stay supported but are not the first-run target.

AX Engine default pack and the rest of the supported catalog (excerpt)
Pack / familyRoleWhy
Qwen 3.8 27B AXQ 6-bit MTP — `qwen3.8-27b:axq`ChatDefault serve pack; qualification SKU
Qwen 3.6 35B-A3B 6-bit MTPChatCoding-focused MoE, multi-model friendly
Qwen3-VL 30B-A3B AXQVision chatQwen3-VL vision language model
Gemma 4 31B 6-bit Assistant-MTPImage / audio / video chatOpen multimodal assistant
Ornith 1.5 35B-A3B coding agentCoding agentQwen3.5-class 35B-A3B MoE

Native multimodal and speech support is checkpoint-authoritative: AX advertises only modalities whose required tower tensors are present in `model-manifest.json`.

Full model matrix, certifications, and MTP pack notes ↗