cd /news/large-language-models/lets-be-realistic-about-kimi-open-so… · home topics large-language-models article
[ARTICLE · art-67338] src=pub.towardsai.net ↗ pub= topic=large-language-models verified=true sentiment=· neutral

Lets Be Realistic About Kimi Open Source

Moonshot's Kimi K3, a 2.8-trillion-parameter open-source model launched July 16, 2026, scores 57 on the Artificial Analysis Intelligence Index, trailing Claude Fable 5 by three points and GPT-5.6 Sol max by two. Despite claims of affordability, K3 costs $15 per million output tokens via API and $0.95 per Intelligence Index task, near parity with GPT-5.6 Sol max at $1.04, before infrastructure costs for local deployment.

read1 min views1 publishedJul 21, 2026
Lets Be Realistic About Kimi Open Source
Image: Pub (auto-discovered)

Member-only story

This article was written after I saw an interview from the DataBricks CEO, which suprised me a lot. Kimi K3 has earned the excitement. Moonshot’s new model lands at 57 on the Artificial Analysis Intelligence Index, only three points behind Claude Fable 5 and two behind GPT-5.6 Sol max in the visible comparison panel. That is a serious result. China has not merely produced another excellent value model. Kimi has put a Chinese model inside the frontier comparison.

The part I want to push back on is the sales pitch that follows: this is supposedly a cheap model that anybody can run locally because it is open. The numbers do not support that conclusion. K3 costs $15 per million output tokens through Kimi’s API. On Artificial Analysis’ completed-task proxy, it costs $0.95 per Intelligence Index task, versus $1.04 for GPT-5.6 Sol max. That is not a price collapse. It is near price parity with Sol, and it is happening before anyone pays for infrastructure, on-call coverage, redundancy, or the hardware required to serve a 2.8-trillion-parameter model.

K3 deserves the headline #

Moonshot announced Kimi K3 on July 16, 2026. The official launch post calls it a 2.8T-parameter model with native vision and a one-million-token context window. It also says the model is designed for long-horizon coding, knowledge work, and…

── more in #large-language-models 4 stories · sorted by recency
── more on @moonshot 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/lets-be-realistic-ab…] indexed:0 read:1min 2026-07-21 ·