spectra-g

SWE 22 years

Projects:

Model Map - bugs - Tools -

[spectra-g @spectragai]

Model Map is still in Alpha looking for feedback 🙏🏻

https://t.co/IGLXBDU5JY

[spectra-g @spectragai]

What a whirlwind few days. NVIDIA they’re just always in the fold

https://t.co/qJrzi9ujfg

[spectra-g @spectragai]

@jjacky What’s the tech stack that handles the automated agent flow?

[spectra-g @spectragai]

@MatthewBerman Router SOLD. Repository SOLD.

What next?

[spectra-g @spectragai]

@lightseekorg @ZixuanLi_ @Zai_org Huge strides made with the new release! And a genius marketing campaign 👏

[spectra-g @spectragai]

How to deploy a website in just 3 words:

⁠npx surge .⁠

That’s literally it and your site is live.

https://t.co/Oi5w0FxXyG is the simplest static site host nobody is talking about. Free SSL, custom domains, agent friendly and no friction!

[spectra-g @spectragai]

GLM 5.3 Flash already at #50 on the OpenRouter 7 day usage rank...it's only been 8 hours!

[spectra-g @spectragai]

@suni_code If you’re referring GLM 5.3 Flash it’s quite cheap? DeepSeek price range

[spectra-g @spectragai]

@araseb_ Lol doing it for the clicks

[spectra-g @spectragai]

@bcherny When will we see a running context length for the main session in Claude Code? It's fundamental to ensuring we are still within the high quality zone.

[spectra-g @spectragai]

@jjacky Or POG style chips

[spectra-g @spectragai]

@jun_song Compute to burn or helped by “friends” at Google?

[spectra-g @spectragai]

@louszbd @jjacky @louszbd did Ox Alpha train/improve on the fly?

[spectra-g @spectragai]

So Ox Alpha was trained on the fly to improve with usage. Has that been done before?

[Lou @louszbd]

Ox Alpha put GLM-5.3-Flash in more people’s hands. We’ve read every piece of feedback. The model you’re using now is better. We will make frontier intelligence more accessible. Enjoy!

[spectra-g @spectragai]

@yngmisu The “friends we made” might possibly have been Google providing the compute

[spectra-g @spectragai]

I think we’re going to see a few free providers for various models surface now, but not necessarily with good intentions.

[spectra-g @spectragai]

@itsPaulAi It’s just above DeepSeek V4 Flash in costs, and rightfully so as it’s just about that in overall intelligence.

These two in particular are in their own league.

I want Kimi to take note and drop its prices significantly.

[spectra-g @spectragai]

GLM 3.5 Flash coding plans start from $18 per month. After the free taster is it enough to jump ship from Claude Code or Codex?

https://t.co/9C7CDlA24a

[spectra-g @spectragai]

Thanks Ox Alpha, the streets won’t forget!

[OpenRouter @OpenRouter]

Ox Alpha revealed: @Zai_org’s GLM-5.3-Flash, the first native multimodal model in the GLM-5 series.

Ox Alpha was the biggest model ever on OpenRouter, processing over 20 trillion tokens in 6 days.

Continue to use the model now: https://t.co/avBrUW8BfZ

[spectra-g @spectragai]

@vipulved @togethercompute Won’t it be revealed by then?