32 lines
1.6 KiB
Markdown
32 lines
1.6 KiB
Markdown
|
---
|
||
|
|
type: xpost
|
||
|
|
source: https://x.com/HealthRanger/status/2078873026989342852
|
||
|
|
author: "@HealthRanger (Mike Adams)"
|
||
|
|
date: 2026-07-19
|
||
|
|
ingested: 2026-07-19
|
||
|
|
tags: [qwen, alibaba, open-source, moe, model-release, china-ai]
|
||
|
|
---
|
||
|
|
|
||
|
|
# Qwen 3.8 — 2.4T Open-Weight Model Announced
|
||
|
|
|
||
|
|
## X-Post (Mike Adams / @HealthRanger)
|
||
|
|
Date: 2026-07-19
|
||
|
|
Engagement: 241 likes, 27 reposts, 18 replies, 27 bookmarks
|
||
|
|
|
||
|
|
> Qwen 3.8 is launching! Right after Kimi-K3, a groundbreaking 2.8T model, Qwen announces their own 2.4T model, which will also be open source (open weights).
|
||
|
|
> The U.S. tech giants are absolutely screwed, and it couldn't happen to a more deserving bunch of evil, greed-driven corporations.
|
||
|
|
|
||
|
|
## Quoted Post (@Alibaba_Qwen)
|
||
|
|
> Qwen3.8 is launching and going open-weight soon!
|
||
|
|
> - 2.4T parameters
|
||
|
|
> - "one of the most powerful model available today, compatible to leading frontier AI models, second only to Fable 5"
|
||
|
|
> - Already available for early testing via Alibaba's Token Plan, Qoder, and QoderWork
|
||
|
|
|
||
|
|
## Context
|
||
|
|
- Second massive open-weight MoE release this week after Kimi-K3 (2.8T)
|
||
|
|
- China's open-source AI output is scaling rapidly
|
||
|
|
- HealthRanger framing is ideological (anti-US-tech), not analytical — the factual signal (2.4T open weights) is significant on its own
|
||
|
|
- "Second only to Fable 5" — unverified claim, Fable 5 not established as standard reference
|
||
|
|
|
||
|
|
## Hector's Assessment
|
||
|
|
Open-source models are compressing margins, yes. But "US tech giants screwed" is overstatement — US frontier models (GPT-5, Claude) retain advantages in reasoning and tool-use. The race is tightening, not over.
|