ainewsblitz.com

Breaking

Ornith Releases Ornith-1.5 Open Models at 397B, 35B, and 9B, Trained via End-to-End Self-Improvement

  • Foundation Models
  • Open Source
  • Software Dev & Coding

Ornith has released Ornith-1.5, a family of MIT-licensed open-weight language models in 9B dense, 35B mixture-of-experts, and 397B MoE sizes, trained so the models propose new tasks, write task-specific scaffolds, and generate solution rollouts for reinforcement learning. The lab says the 397B flagship is on par with Claude Opus 4.8 on several agentic coding benchmarks and leads open models of comparable scale, with full-precision and quantized weights posted on Hugging Face.

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 27,568 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year