Ornith-1.5: New Open-Source AI Rivals Claude Opus 4.8

Ornith has announced Ornith-1.5, a new open-source model family. It includes three coding models that Ornith claims have benchmark scores putting its largest system near Claude Opus 4.8. Here's the thing: The models are open and the license is permissive, which is a practical hook for developers. Ornith-1.5 includes a 9B dense model, a 35B mixture-of-experts model, and a 397B MoE flagship. These are released under the MIT license, with quantized versions available. Ornith states the 397B version scored 86.1 on Terminal-Bench 2.1 and other high marks on various benchmarks. What's interesting is that these numbers are vendor-reported. While impressive, independent verification is still needed. For comparison, Claude Code with Opus 4.8 scores 78.9% on Terminal-Bench 2.1 on the public leaderboard. The bottom line is that if these benchmark claims are reproduced, Ornith-1.5 could offer a powerful, open-source alternative for agentic coding systems.