Qwen3.8 27B

Qwen3.8 27B is a large language model entity within the Qwen series. It serves as the underlying engine for various agentic frameworks and demonstrations.

Integrations & Demos

• Hermes Agent Framework: The model is utilized in hermes-agent, an open-source AI agentic framework developed by nous-research.
• /loop Feature: Demonstrated in the context of automated recurring tasks.
  • The framework supports a /loop command for scheduling and executing recurring operations.
  • See loop Feature: Automated Recurring Task Demo for a hands-on demonstration using Qwen3.8 27B.

Performance & Quantization

• Swift 1.5 Optimization: The model is available in the quantized Swift 1.5 variant (GSQ-RCO IQ3_S), optimized for local deployment on hardware with limited VRAM (e.g., 16GB setups).
• Benchmarking: Comprehensive performance evaluations for the Swift 1.5 Qwen3.8-27B-GSQ-RCO-GGUF model are documented in Swift 1.5 Qwen3.8-27B GSQ-RCO IQ3_S 16GB LLM Performance Benchmark.
  • Tests cover performance metrics on local Ubuntu server setups with RTX 2000 Ada GPUs.

References

• Hermes Agent’s New /loop Feature: Automated Recurring Task Demo
• Swift 1.5 Qwen3.8-27B GSQ-RCO IQ3_S 16GB LLM Performance Benchmark