Composer 2 Sparks Kimi Attribution Row

The optics are rough: if a “new” in-house model is really an open-weights base with RL on top, the only thing that matters is how honestly Cursor explains that stack.

• –Cursor’s official 2.0 launch framed Composer as its first agentic coding model, so provenance questions cut straight through the product story.
• –The alleged Kimi connection matters less for raw capability than for licensing, attribution, and whether premium pricing matches the actual technical lift.
• –This is the kind of controversy that can spread fast in AI coding circles because builders care about trust, not just benchmark claims.
• –Even if the underlying model is strong, vague sourcing turns a release win into a credibility tax.
• –The broader takeaway: in 2026, “proprietary” AI branding gets scrutinized almost instantly, especially when the community can inspect the traces.