@WR4NYGov @mattshumer_ Not even 5M can match frontier fable/gpt 5.6 intelligence.
Local AI is cope.
Unless it’s your hobby, then its hella fun.
English
🦉
61 posts
I’d like to play
Local AI models can be fun to play with but are not great for serious work, unless …
You’re big enough to spend over $100K on hardware and run it more than about 12 hours a day.
Not sure about the $100K. Might be $5M.
Please tell me how I did and explain what I got wrong.
English
Getting rate limited is fine
Killing the task halfway through is not
claude code should just let the task finish like codex does
English
muse spark 1.1 is #3 on Debate Benchmark, only behind Fable 5 and Opus 4.7 and ahead of GPT-5.6 Sol
Lech Mazur@LechMazur
GPT-5.6 Sol (1567 → 1684), Grok 4.5 (1410 → 1524), Sonnet 5 (1604 → 1622), MiniMax-M3 (1481 → 1541) all improve on their predecessors' scores in the Debate Benchmark. But the big surprise is Muse Spark 1.1: it debuts at 1688, 3rd overall behind only Fable 5 and Opus 4.7.
English
We care deeply about your privacy and respect customer choice. For teams using zero data retention, no trace and code data is ever retained. All API key use of Grok Build also respects ZDR.
If ZDR is disabled, the /privacy command is available in the CLI to disable data retention, which also deletes previously synced data.
Run the /privacy command to view or change your settings at any time.
English
English
@ns123abc i don’t understand, tibo literally said they reverted it? what are you talking about
English
@NoOnemklz @jun_song GLM 5.2 isn’t even in the same league. Don’t know why it keeps getting mentioned, it’s open weight, cheap and competitive; but not as a frontier model.
English
@jun_song Here Grok 4.5 crushes GLM 5.2 hands down in performance!
Lisan al Gaib@scaling01
"GLM-5.2 is Mythos-level" 🤡🤡🤡
English
Just for the information, using Grok 4.5 via API (in OpenCode) empties limits faster than using it via Grok Build CLI?
English
@Tubbymcwub @peer_rich Or they wait it out untill the inference becomes more efficient and cost effective
English
the correction will hit worse than the $10 uber drives
Corey Haines@coreyhainesco
I'm paying $200 for $8,545.93 in usage. If you don't think the AI subsidization apocalypse is coming, you're mistaken.
English
Free for everyone (outside EU 😢)
Polymarket@Polymarket
JUST IN: SpaceXAI makes Grok 4.5 free to try for users in Grok Build CLI & Cursor.
English
Try Grok 4.5 for free in Grok Build!
curl -fsSL x.ai/cli/install.sh | bash
Grok@grok
Grok 4.5 is now available to try on the free tier. Use Grok Build with any X or Grok account. We’re excited to hear your feedback. x.ai/cli
English
Fable 5 release: 90% of X talking about it.
GPT-5.6 release: About 25% of X talks about it.
Grok 4.5 release: Literally 2 people in my feed 😄.
Clear levels of trust in the models. While we may not agree with Anthropic's rollout approach (and rug pull not their fault entirely), they still have consumer trust over the rest. Opus model earned them immense customer retention.
Not surprised at all by quietness of Grok 4.5. I've tried other Grok models a few times and it feels like 2024 AI. Gonna need a lot of positive results to earn that trust back.
English
@Panchoavila @saltyAom Ngl grok 4.5 high has been surprising me fixing shit gpt 5.5 pro xhigh created
English
One day of heavy Grok 4.5 usage: 1%.
One day of heavy Fable 5 usage: 100% of your weekly limit.
I have been hammering Grok 4.5 inside Grok Build all day on a SuperGrok subscription.
The weekly meter says 1%.
At this pace I could build every single day, all week, on ONE subscription.
A concept Anthropic has apparently never considered.
English