on:qwen-3-8-flash-next

· 10 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

• @xlr8harder 2026-08-30 — it's kind of funny that all of the distillation means that the chinese models are adopting the worst of opus's behaviora ♥126
• @witcheer 2026-09-09 — a model 2.5x bigger than my GPU just posted the highest MMLU on my RTX 5090 board. Qwen3.8-Flash-Next. 125B parameters, ♥108
• @__venki__ 2026-09-17 — Have a task that wants cheap text processing, and I’m trying all the cheap models. (Glm 5.3 flash, qwen 3.8 flash) They ♥81
• @__venki__ 2026-09-18 — Conclusion from my experiments: Tested: Deepseek V4.1 flash, Qwen 3.8 flash, glm 5.3 flash, muse spark, gpt Luna Rough ♥8
• @geogristle 2026-09-03 — @k7agar qwen flash next on 2x 3090 @ 250W, trying to grab a third gpu today ♥5
• @jmbollenbacher 2026-08-26 — After seeing the architectures glm-5.3-flash and qwen-38-flash-next, i think maybe this take was wrong. We're going to ♥4
• @jmbollenbacher 2026-08-26 — glm-5.3-flash and qwen3.8-flash-next in the same day. what an embarrassment of riches. ♥3
• @multimodalart 2026-08-26 — it's great that with Qwen3.8-Flash-Next and GLM-5.3-Flash, we got not only 2 high quality models, but also 2 novel model ♥2
• @normie_dot_exe 2026-10-19 — Qwen 3.8 Flash makes reasonable plans, and communicates well, but struggles with game knowledge https://t.co/i9Cgvfo8fQ ♥1
• @RealEverNever 2026-08-28 — @QwenDevs Starting with the Unsloth GGUF. Local Qwen is on characters this week. I need to know if 6B active can hold a ♥0