REDDIT · REDDIT// 3h agoMODEL RELEASE

Qwen3.6-35B-A3B hits instruction-following wall

Qwen3.6-35B-A3B, a sparse Mixture-of-Experts (MoE) model released in mid-April 2026, is facing community backlash for failing to adhere to system instructions and context files. Despite high SWE-bench scores, developers report significant regression in agentic obedience and tool-calling reliability compared to older models.

// ANALYSIS

The transition to sparse MoE appears to have traded reliable instruction adherence for benchmark performance. Reasoning loops and "thinking" mode bloat frequently lead the model to ignore constraints, while a high tool-call error rate and technical instability with FP8 KV cache make it a risky choice for production agents compared to Qwen 3.5.

// TAGS

qwen3.6-35b-a3bllmagentreasoningprompt-engineeringopen-weights

DISCOVERED

3h ago

2026-04-23

PUBLISHED

4h ago

2026-04-23

RELEVANCE

9/ 10

AUTHOR

FusionX