liuwen

back to the future

[liuwen @liuwen]

@mxfp4 why not simply cliproxycli. easy to extend with plugins or even easy enough to fork your own small hacks into scheduling policy if you like. I did that before

[liuwen @liuwen]

@thorstenball when I saw your post snippet in the feed I thought you were joking or a big 'though' later... oh bro in China you get everything

[liuwen @liuwen]

@thsottiaux @grok what's your take in this minor drama

[liuwen @liuwen]

@sqs the name is `amz` thanks!

[liuwen @liuwen]

@thesammykins @AmpCode you mean like `creative writing`? I found gemini models used to be nice with this last year, but not superior anymore sadly. anyway, isn't gemini-3.6-flash is cheap enough?

[Malte Ubl @cramforce]

We ran Kimi K3 on a private cybersecurity benchmark.

TL;DR: Kimi K3 is the workhorse for cyber security tasks at great recall/precision/price. GPT 5.6 is best recall/precision but at 7x higher cost per run.

For context, https://t.co/UMvysvNW5w is an open-source cyber harness designed for finding vulnerabilities in large codebases.

The eval runs deepsec on an undisclosed open-core application at a git sha before a large number of security issues were fixed. This is a secret eval that cannot be directly benchmark-maxxed.

S-Tier: GPT 5.6 Sol: By far the most thorough analysis, but coming in at over 7x the price of the runner up.

Best price/recall: Kimi K3. Next tier of recall at a good price

Best price at good recall: GLM 5.2 (40% lower price than Kimi K3)

GPT 5.5: Only recommended with subscription or high-discount API price. Similar recall to Kimi at much higher list price.

Opus 4.8: Only recommended with subscription or high-discount API price. Similar recall to GLM 5.2 at much higher list price.

Fable 5: 100% refusal rate. Cannot be used for security analysis.

Sol on a large code base will quickly get into 6-figure pricing. This is still affordable relative to the risk of letting security issues unfixed or paying bug bounties.

I'd recommend using Sol for a one-time baseline and then using Kimi K3 for continuous analysis.

When using open-weight models, make sure to use an inference vendor that supports zero data retention.

[liuwen @liuwen]

@Iberia_en Thanks, but could you please try that number yourself? [phone] is definitely a dummy.

[liuwen @liuwen]

first tweet from a mobile device, finally

[liuwen @liuwen]

wandering