One memory. Every model. Always private.
San Francisco
- Tweets 1,585
- Following 9
- Followers 19,916
- Likes 347
Anuma retweeted
im 25
i’m building an ai social discovery app
single dad to a beautiful dog named Frog 🐸
born and raised in jersey city. refuse to move to SF.
my longest streak is 8 continuous years on grindr
happy to connect with other founders and out of pocket people!
5
1
8
185
Anuma retweeted
I’m 23
Korean, based in sf, currently in nyc for launch
Berkeley econ & stats
Founding growth & ops for an ai social discovery app
New to X, looking to connect with other builders :)
3
1
8
179
GPT 6 Sol and Luna are live in Anuma.
Run them next to Claude Opus 5.5, Fable 5.1, Grok and Kimi. Same memory, same chat.
Available now.
Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe.
GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale.
We’ve also made caching and inference more efficient, and we’re passing the savings directly to you: 50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing.
1
2
5
660
GPT-6 Sol and GPT-6 Luna are now in Anuma. ⚡️
GPT-6 Sol: Built for deep reasoning, complex problems, coding, and long-running work.
GPT-6 Luna: Fast and efficient for everyday questions, writing, analysis, and more.
Try them in Anuma today: chat.anuma.ai
1
1
4
496
👋 Anuma is looking for creators!
We’re building out our Creator Program and looking for TikTok/Instagram creators who want to get paid to post – no huge following required, just good content instincts and consistency.
📱 1 TikTok + 1 IG Reel per day
💰 Up to $140/week
💡 We provide the hooks, trends, and messaging – you bring your voice
🎥 Your phone is all you need
If you know someone who’s great at short-form video (or that’s you 👀), this could be a great fit.
Apply here → forms.gle/WyacnCqGNbmyEVk79
#Hiring #CreatorEconomy #UGC #ContentCreator #SocialMediaJobs #TikTok #Instagram
6
3
8
651
GLM-5.3 is developed by Z.ai @Zai_org, the team behind the GLM family of AI models.
What makes it interesting?
→ 1M-token context
→ Multimodal
→ Strong coding capabilities
→ Built for agentic tasks
→ Designed for long-running, complex workflows
In plain English: GLM-5.3 is built to do more than answer questions. Give it a complicated task and it can reason through the steps, use tools, write code, and keep working toward the goal.
ALT Screenshot of AI interface with model options and features on orange background.
3
2
6
613
Kimi K3 vs GLM-5.3
Both are built for complex, long-running AI work.
Kimi K3 Strength
🧠 Long-horizon reasoning
👀 Native vision
📚 Large documents + knowledge work
💻 Coding + research
GLM-5.3 Strength
💻 Coding + software engineering
🤖 Agentic workflows
🔧 Tool use + automation
🧩 Complex technical tasks
Think of it this way:
Kimi K3 → understand and reason through a big problem
GLM-5.3 → work through a complex task and get things done
Different models. Different strengths. That's why having more than one model matters.
1
1
181
GLM-5.3 Fun fact
Z.ai says GLM-5.3 has become an “indispensable daily coding partner” for its own team.
And GLM-5.3 even helped build the infrastructure used to run its successor, GLM-5.3-Flash.
GLM-5.3 is included in Anuma. Try it and see what you can build: chat.anuma.ai
46
How well do you actually know about all the AI models within Anuma? GPT, Claude, Gemini, Grok, Kimi, Qwen ...
What’s the difference between them? What is each one good at?
Let's learn about 1 AI model every day.
Day 1: Kimi K3 👇
ALT Chat interface with model selection dropdown, on orange gradient background.
2
3
9
959
🧠 Kimi K3 @Kimi_Moonshot
Who made it?
Kimi is developed by Moonshot AI, an AI company based in China.
What makes K3 interesting?
→ 2.8T parameters
→ Up to 1M token context
→ Native vision
→ Reasoning + coding
→ Built for long, complex tasks
→ Open weights
In plain English: K3 is designed to handle lots of information and work through tasks that take multiple steps.
1
1
556
Fun fact: Kimi K3 is the first open model at the 3 trillion parameter scale.
So what does that actually mean for you?
Try giving K3 a long document, a big research project, or a complicated task and see how it handles it.
Kimi K3 is available in Anuma. Try it: chat.anuma.ai
1
2
2
80
What if every AI could remember you without owning your data?
That’s Anuma.
Open source AIs get memory.
Closed source AIs get more privacy.
You keep control of both.
One private memory layer for all your AI.
ALT Comparison chart of AI memory features for Anuma, ChatGPT, Claude, and Gemini.
2
2
8
808
most people have no idea why they'd ever pick an open-source model over gpt or claude
short version:
open = nothing stored, nothing logged, nothing trained on
closed = more raw capability
both are in every anuma plan. pick per task, not per subscription 🧡
ALT Comparison table of open-source and closed-source LLMs with details.
1
2
6
573
Meta's Muse can remember what matters to you and act on your behalf.
It makes us wonder: What else can a personal AI do with everything it knows about you? other than,
Book your trip? Send your emails? Buy your groceries?
What about:
I think you’d really get along with this person.
#muse
ALT Illustration of AI integration with farm elements, arrows, and text 'AI Knows You. Now Can It Find Your People?'
2
2
6
502