Hacker News Reader

Comment by felixgallo | Hacker News Reader

[flagged]

I am because Claude has figured out I am a biochemist and therefore even asking Fable what the weather is gets me bumped back to Opus, sometimes even Opus 4.8 instead of 5. Kimi? GLM 5.3? DeepSeek? No such problem.

I can literally open a new chat with just "Hello" and it gets bumped.

Memories break Fable 5 for me as well in chatbot. I ask Opus a lot of sec related stuff and now if I even type “hello” in chat it gets insta-downgraded to Opus 5.

Wow, that's totally crazy. Does it happen even if you clear out the "offending" memory entries?

Just cancelled my Claude subscription and used Ox Alpha Free through OpenCode (so that's one).

In my view, from the testing I did since Thursday, it's better than Fable. I had just finished a rather large task that Fable completed, including a /review and an /ultrareview.

Ox Alpha found bugs that Fable and Opus missed, and it continued to build things like a pro.

It does have issues with availability - but it's on a free promo right now. That also means that I don't know how much it would have cost if I had to pay API prices for it, which might not be cheaper than the subsidized small company / consumer usage, but for large companies paying API prices for either offering, the difference will be considerable.

Intriguing. Who is it? What is it? any ideas?

https://openrouter.ai/stealth/ox-alpha

Some people say it is a new version of Gemini Pro - this is based on some tweets from their employees.

Rumored to be mimo

Are you sure, most indicators suggest Z.ai with a potentially Flash or Air variant of GLM 5.3?

I almost feel like there's some sort of paid campaign going on in Hacker News promoting the Chinese/open models. I feel like every day I'm hearing about how the frontier labs are dead but your experience is the same as mine. My company pays for Claude AND Codex but we never really use the open models for anything critical.

Doesn’t matter if your company pays for Claude, Anthropic and OpenAI valuations and expenditures commitment requires them to win the vast majority of the market to make economical sense. And the Chinese competition makes that very unlikely, to say the least. They likely won’t disappear fully but their “free” lunch as the AI darlings is done, on paper. Will be interesting to see how they adapt

Sure, I'm being "paid" by the fact that it costs pennies a day to use Deepseek and if I could afford the hardware then I could run it locally. Meanwhile, Anthropic is obviously a threat to open weight models and actively lobbies the US Government to have them banned or controlled.

So yeah, I and I guess others, are quite active in whatever little way we have available, to up vote new models and share stories.

Reply on news.ycombinator.com