OrcaRouter
2026OrcaRouteropen source
Qwen3.8-27B-Uncensored-MLX
abliterated
uncensored
Qwen3.8-27B-Uncensored-MLX is an abliterated, refusal-removed version of the Qwen3.8-27B vision-language model, specifically optimized for Apple Silicon using the MLX framework. It is designed for AI safety research, red-teaming, and interpretability studies, providing full access to the model's capabilities without built-in guardrails.
Abliterated (refusal-removed) safety alignment for research
Native vision-language capabilities with MTP head
Optimized for Apple Silicon via MLX framework
pricingFree (Open Source)
regionUnited States
2026OrcaRouteropen source
OrcaSAQ-2 27B
v2
quantization
reasoning
OrcaSAQ-2 27B is a highly compressed, sensitivity-aware mixed-precision quantized version of the Qwen3.8-27B model. It reduces the storage footprint by over 77% while maintaining high fidelity to the original model, making it suitable for deployment on hardware with limited GPU memory.
3-bit sensitivity-aware mixed-precision quantization
Supports MTP speculative decoding and tool calling
Optimized for long-horizon agents and complex reasoning
pricingOpen weight (Apache-2.0 license)
version2