Qwen3.8-Flash-Next gains Strix Halo support on Strata

A developer released official Strix Halo support for Qwen3.8-Flash-Next on Strata, claiming best long-context decode performance, up to 1M context length, Linux-only and marked experimental.

• User Shares Strata Speedup for Local LLMs but 64GB RAM Constraint2026-10-04
• Qwen3.8-Flash-Next Hits 65 Tokens/Second on 12GB VRAM2026-09-24
• Qwen3.8-Flash-Next Hits 1M Context on Strix Halo2026-09-20
• Strata accused of rewriting GitHub history to hide Claude authorship2026-10-09
• User builds 256GB VRAM rig with 8 Radeon Pro V620s2026-10-08