고정된 트윗
naeem (e/acc)
1.4K posts
@itsOmSarraf_ Isn't as good as cursor glass, which is basically their version codex app
English
@natolambert Fair, as long as they don't violate licensing of somebody else's work.
English
My take on the cursor-kimi matter is that cursor is somewhat like openai and they just think about themselves and how exciting their model is and that's actually fine.
English
@bridgemindai K2.5 is a multimodal.
Elo score has nothing to do with the base model quality. They have probably evaluated every major open weights base model in order to come to this conclusion.
Composer 2 is a great model.
English
Cursor built Composer 2 on top of Kimi K2.5.
Kimi K2.5 ranks #14 on LMArena Code with 1431 Elo.
Behind Claude Opus 4.6. Behind Claude Sonnet 4.6.
Behind GPT 5.4. Behind Gemini 3.1 Pro. Behind GLM-5.
Behind MiniMax M2.7.
You're telling me Cursor picked the #14 ranked coding model as their base?
MiniMax M2.7 is #9. GLM 5 is #8.
Both cheaper.
Both better at code.
Now you understand why Composer 2 hallucinates all over the place. The foundation wasn't strong enough.
English
@aiamblichus @XiaomiMiMo Yes, a coding plan would be greatt
English
@XiaomiMiMo It's a great model. Are you thinking of offering a coding plan?
English
it is our MiMo-V2-Omni :)
αιamblichus@aiamblichus
I don't know who's behind the "healer-alpha" model on OpenRouter but: - it has replaced 95% of my Sonnet usage - it's omni-modal - it's fast and furious, esp inside pi - it's probably Chinese If this is open-sourced and/or priced competitively, this will be a *big deal*
English
This is happening!
GIF
Pavan Davuluri@pavandavuluri
The team and I have spent the past several months analyzing feedback from the community. What came through was the voice of people who care deeply about Windows and want it to be better. Read this blog post to learn more about what we're doing in response as we look to raise the bar on Windows 11 quality. Please keep the feedback coming, to help us shape the future of Windows together. blogs.windows.com/windows-inside…
English
@iamdamianb They probably have Early access to whatever next version slopus is, they chillin'
English
I feel bad for the engineers at Anthropic.
They cant use GPT 5.4.
English
English
@singularvector @Jarsen @mjackson No, you can’t: you can use GPT, their model family, in OpenCode… Codex is an alternative harness to OpenCode.
English
Agentic capability may be our remaining shortcoming due to some legacy issues, but we are putting all our efforts into improvement. Stay tuned for our top-tier agentic foundation models!
Arena.ai@arena
Qwen 3.5 Max Preview has landed in top 10 for Arena Expert and top 15 for Text Arena. It shows particular strength in Math. Highlights: - #3 Math - #10 Expert - #15 Text Arena - Top 20 for Writing, Literature & Language, Life, Physical, & Social Science, Entertainment, Sports, & Media, and Medicine & Healthcare Congrats to the @Alibaba_Qwen team for this new milestone!
English
@zxytim Aak dario for lawyers, I heard he loves oss models
English
@jpschroeder I am tired of having to switch to opus for the frontend only.
English
The good news is Codex is way better than Opus at literally everything except UI
dax@thdxr
opencode 1.3.0 will no longer autoload the claude max plugin we did our best to convince anthropic to support developer choice but they sent lawyers it's your right to access services however you wish but it is also their right to block whoever they want we can't maintain an official plugin so it's been removed from github and marked deprecated on npm appreciate our partners at openai, github and gitlab who are going the other direction and supporting developer freedom
English
The best codex feature is the model... Get it. Very impressive.
English
@thsottiaux Make it better at frontend and it will be flawless.
English
T3 Code now supports Claude.
If you have the Claude Code CLI installed and signed in, you can use it with T3 Code.
Hopefully the lawyers won't make us remove this 🙃
English
@AndresDevvv @SteveDun @theo @jetpackjoe_ Not everybody wants to create a million dollar saas using claude opus 4.6 max reasoning bro, sometimes people just want to use it to the potential which every size class of oss models provide.
English
@singularvector @SteveDun @theo @jetpackjoe_ Nobody is ever getting serious work done on local models bro
English
Who’s got a polymarket on time till we get a legal notice?
Theo - t3.gg@theo
T3 Code now supports Claude. If you have the Claude Code CLI installed and signed in, you can use it with T3 Code. Hopefully the lawyers won't make us remove this 🙃
English
English
