Kimi K3 benchmark chart puts it near the top on coding tests, with B300 jokes

MaziyarPanahi · x · 2026-07-27

A follow-up reaction to Kimi K3 says the model “needs 8x B300,” implying heavy compute requirements.

The attached benchmark image shows Kimi K3 near the top on several coding-related tests, including Terminal Bench 2.1, Program Bench, and SWE Marathon, while other models like GPT-5.6, Fable 5, Opus-4.8, and GLM-5.2 are listed for comparison.

Original post →

More from Models

Models channel →