ainewsblitz.com

Breaking

Tencent Releases Hy3, a 295B Mixture-of-Experts Model That Runs on Just 21B Active Parameters

Tencent Releases Hy3, a 295B Mixture-of-Experts Model That Runs on Just 21B Active Parameters
  • Foundation Models
  • Open Source
  • AI Agents

Tencent has officially released Hy3, a 295-billion-parameter Mixture-of-Experts (MoE) model that activates only 21 billion parameters per token, positioning it as a cost-efficient open-weight rival to flagship models several times its size. The model, part of Tencent's Hunyuan (Hy) series, launched on July 6, 2026 under the permissive Apache 2.0 license and was distributed through Hugging Face, GitHub, and ModelScope, where it drew more than 20,000 downloads shortly after release.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,934 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year