Qwen 3.8-Flash-Next Claimed to Match V4-Flash Tier at ~4x Smaller Size

teortaxesTex · x · 2026-09-11

A user notes Qwen 3.8-Flash-Next reportedly sits in the same Artificial Analysis tier as V4-Flash, despite being almost 4x smaller with nearly 3x fewer active parameters. The poster admits limited firsthand use but argues that if the comparison holds, the model is significantly underhyped — an unverified capability comparison worth watching.

Original post →

More from Models

Models channel →