Kimi K3-256k

This was posted 38 minutes ago, and as of 20 minutes ago, several Anthropic services are now designated as having a "major outage".

Doubt these are related, but it made me laugh a little.

My current Claude Code session has been going on for like 35 hours and has used up around 400 million tokens, thankfully almost all of those being cached (95-98%) - pretty typical for long form agentic work.

First you spend like 2-3 hours working on a plan, once you have that you just tell the model to go and implement it, do adversarial sub-agent review loops before each commit and also make sure that all tooling and tests pass (including coverage requirements). You do need to poke it in a slightly different direction every few hours, though. Not even any novel work, just some refactoring and SSE notification hardening, bug fixes, alongside environment tuning and getting rid of some bottlenecks (also migrated from Oracle to PostgreSQL but that's mostly done).

That said, Kimi somehow manages to use less context in the main thread than Anthropic's models (even when you use sub-agents and also dynamic workflows in Claude Code), might have something to do with either how the model is tuned or their Kimi Code harness - because even in most of the longer form sessions it doesn't seem to fill up quite as quickly (note: because the kimi vis tool doesn't have a full summary view across all agents, these are the main long running agent stats across some sessions, not sub-agents):

  total tokens    cache hit rate    wall time    peak context
  283M            98%               3963m        466k
  258M            97%               2724m        467k
  98M             94%               1353m        393k
  67M             97%               614m         434k
  75M             98%               1447m        498k
  53M             99%               191m         375k
  6M              96%               139m         124k
  7M              98%               86m          118k
  11M             99%               61m          147k

I could see 256k context being sufficient for all sorts of work, even if intermediate progress/plan tracking files and docs might have to be used along the way, in addition to whatever plan support the harness has (for example, if you document something that will be relevant for load testing you might need that in 10 turns but not during the ones before then).

> Are they really that good,

They did exceed the expecations of pretty much everyone! I've also blogged about using the model, it's a bit on the slow side but pretty good!

> so were totally surprised and overwhelmed by the requests ... or do they just don't have the hardware being in china?

Yes, this is mostly the case: https://x.com/Kimi_Moonshot/status/2078855608565207130

As a user, I much prefer that to service disruptions or severely degraded or secretly quantized performance. However if I didn't have an account, I'd be pretty pissed off about not being able to give them money and become a user.

No, the waiting list is true.

Kimi had become that popular. I was a subscriber of Kimi back when latest version was Kimi K2. Later I unsubscribed because I jumped over to GLM subscription (they had amazing deal). Now when I wanted to try out Kimi K3 to find out what the fuzz was all about, I couldn’t subscribe to them.

I remember reading a post from Moonshot team about this, they are doing this because they are almost at peak capacity and want to reserve it to keep the quality for their current customers.

We are actually witnessing an open-weight model catching up at catching mainstream users attention. And instead of behaving like Anthropic, they actually care about their users experience.

The only downside with third party providers is that you have to trust that the provider have setup and configured it correctly, and is not secretly quantizing it.

See Kimi Vendor Verifier.