K3 Reportedly Outperforms Opus 4.8 Across the Board

basedjensen · x · 2026-07-17

A repost claims that some benchmark scores for the new K3 model have been officially confirmed.

The key takeaway is that this is a Fable/Sol-level model, reportedly outperforming Opus 4.8 across multiple benchmarks, yet priced at the Sonnet tier. The poster describes this release as the most impactful model update since DeepSeek R1.

Related event: Kimi K3 Debuts Strong, Narrowing the Open-Weight Gap(184 posts)→

Original post →

More from Models

Models channel →