Kimi K3 is now available in ThinkReview

Kimi K3 is now available in ThinkReview

We're excited to announce that Kimi K3 from Moonshot AI is now available in ThinkReview. In their Kimi K3 tech blog, Moonshot introduces K3 as their most capable model yet — a 2.8-trillion-parameter, open 3T-class model with native vision and a 1-million-token context window. You can use it today for AI-powered code reviews on GitHub, GitLab, Azure DevOps, and Bitbucket.

What is Kimi K3?

Kimi K3 is Moonshot's open frontier model for long-horizon coding, knowledge work, and reasoning. It builds on Kimi Delta Attention and Attention Residuals, with a highly sparse Mixture-of-Experts design that activates 16 of 896 experts. Compared with Kimi K2, Moonshot reports roughly 2.5× better scaling efficiency — more intelligence per unit of compute.

Highlights from the official announcement:

• First open 3T-class model — 2.8T parameters at the open scaling frontier.
• Built for long-horizon coding — Sustains long engineering sessions, navigates large repositories, and orchestrates terminal tools with minimal oversight.
• Vision in the loop — Native multimodal architecture that reasons over screenshots and visuals for frontend, game, and UI-adjacent work.
• 1M-token context — Room for large diffs, related files, and deep repository context when ThinkReview can fetch it.
• Competitive coding benchmarks — Strong results across DeepSWE, Terminal-Bench, SWE Marathon, FrontierSWE, and related agentic coding suites — trailing only the strongest proprietary flagships in Moonshot's evaluation.

K3 is available on Kimi.com, Kimi Work, Kimi Code, and the Kimi API — and now inside ThinkReview for everyday pull request review.

Why Kimi K3 is strong for code reviews

Code review is exactly the workload K3 is tuned for: sustained reasoning over large codebases, careful multi-file analysis, and actionable engineering feedback.

• Long-horizon PR analysis — K3 is designed to stay coherent across big, multi-step engineering tasks — the same muscle you need when a merge request spans services, configs, and tests.
• Repository-scale context — A million-token window helps keep large diffs and surrounding code in scope instead of truncating the parts that explain why a change is risky.
• Vision + code when it matters — When a PR includes UI screenshots, diagrams, or visual regressions, K3 can reason across image and code in one pass.
• Open frontier quality — Frontier-level coding performance without locking your review stack to a single closed vendor.
• Practical on messy brownfield work — Moonshot showcases K3 on kernel optimization, compiler work, and research-grade pipelines — the same class of careful, mechanism-level reasoning that catches race conditions, API contract breaks, and subtle regressions in real PRs.

Pair K3 with repository-level context when the real bug lives outside the hunk, and treat severity labels as input — your team's bar for "critical" still wins.

How to use Kimi K3 in ThinkReview

1. Open ThinkReview settings — Click the extension icon and go to Settings or Model selection.
2. Choose Kimi K3 — Select it from the model dropdown for your reviews.
3. Run a review — Open any pull request or merge request and start ThinkReview with your chosen model.

Kimi K3 is available to Professional and Teams plan subscribers. Manage your catalog in Model Selection on the ThinkReview portal.

Where it works

Same as always: GitHub, GitLab, Azure DevOps, and Bitbucket Cloud. One extension, one workflow — now with Moonshot's open frontier model for deep AI reviews.

If you try Kimi K3 on real PRs and have feedback, we'd love to hear it — open an issue on GitHub or reach out via thinkreview.dev.

Ready for open frontier reviews on your next PR? Install ThinkReview or manage your models in the portal.

Model details reference Moonshot AI's Kimi K3 announcement.