Agents
For real agentic workloads (North), short-context calibration wasn't enough. We calibrated AWQ on long internal agentic traces (up to 64k to…
For real agentic workloads (North), short-context calibration wasn't enough. We calibrated AWQ on long internal agentic traces (up to 64k tokens) and added token masking in llm-compressor to exclude r
For real agentic workloads (North), short-context calibration wasn't enough. We calibrated AWQ on long internal agentic traces (up to 64k tokens) and added token masking in llm-compressor to exclude repetitive chat templates/tool descriptions from calibration stats. Plus QAD (quant-aware distillation) to close the last gap — matching the quality of our BF16 MoE model with W4A8.
Related
- APEX-MEM: Agentic Semi-Structured Memory with Temporal Reasoning for Long-Term Conversational AI
- GenericAgent: A Token-Efficient Self-Evolving LLM Agent via Contextual Information Density Maximization (V1.0)
- Introducing Kimi K2.6 from @Kimi_Moonshot, a multimodal agentic model with Agent Swarm scaling to 300 sub-agents and long-horizon coding sta…
- When Less Latent Leads to Better Relay: Information-Preserving Compression for Latent Multi-Agent LLM Collaboration
- HeLa-Mem: Hebbian Learning and Associative Memory for LLM Agents
Source: Cohere (X) | 2026-04-22