AI Trend Notifier
EN

$ diff granite-4-2 kimi-k3

Granite 4.2 vs Kimi K3

Values come from Granite 4.2 and Kimi K3, where each is cited to its source. This page states no benchmark result and ranks neither model — it puts two published specifications next to each other. Where a lab has not published a figure, the row says so rather than guessing.

What actually differs

Context window
Kimi K3 takes 1M against 512K — modestly more room in a single request.
Weights
Both publish weights — Granite 4.2 under Apache 2.0, Kimi K3 under Modified MIT (commercial use permitted; weights publicly downloadable). Check the licences rather than assuming they permit the same commercial use.
Recency
Granite 4.2 shipped 40 days after Kimi K3 (2026-08-25 vs 2026-07-16).

Full spec

AttributeGranite 4.2Kimi K3
DeveloperIBMMoonshot AI
Released2026-08-252026-07-16
Context window512K (30B); unknown (3B, 8B)1M tokens
Pricingnot applicable — open weights, no hosted rate published$0.30/M cache-hit input · $3.00/M cache-miss input · $15.00/M output
LicenseApache 2.0Modified MIT (commercial use permitted; weights publicly downloadable)
AvailabilityHugging Face, Ollama, GitHubKimi Code, Kimi app, Kimi API; weights on Hugging Face (`moonshot-ai/kimi-k3`, released 2026-07-27)

How these pages are produced

← all comparisons