Want to help build a more regenerative and ethical AI future? @openbmb ambassador
Australia
Joined January 2026
- Tweets3.4K
- Following216
- Followers360
- Likes2.2K
Pinned Tweet
Haven't received a trophy for a long time! MiniCPM ambassador trophy received 🫡
Appreciate the sentiment and the acknowledgement from the @OpenBMB team.
I look forward to building more products that help with my mission to build AI for the good of humanity and the world.
Small models are the way of the future and I'm glad there are labs like OpenBMB who make them a priority.
Love to see what the community builds with these small language models.
When is the wave of affordable Chinese VRAM coming. I know they're talking about DDR ram by 2027 but I want VRAM and I want open weight models that have been tuned to non-CUDA infrastructure.
Let's bring down the monopoly.
en.mthreads.com/
2
86
Josh retweeted
ERNIE 5.5 Flash Preview scored 1668 on the latest Code Arena WebDev leaderboard, ranking #2 among open-source models and #4 among Chinese models.
This is a good step forward, and we’ll keep working to improve the model’s coding capabilities.
Ernie 5.5 Flash Preview by @Baidu_Inc is trending towards becoming the #2 open model on the Code Arena: WebDev, at 1668 pts. The weights will be open-sourced under Apache 2.0 soon.
Its score puts it just above:
- DeepSeek-V4.1-Flash (Max) at 1664
- Kimi K3 (Max) at 1658
and on par with:
- Qwen3.8-Max at 1670
Congrats to the @Baidu_Inc team on this release!
Josh retweeted
Small model showdown incoming.
@Krafton_AI Raon-1 nano launch chart puts it ahead of Qwen3.5 2B, Gemma 4 E2B, @OpenBMB MiniCPM5 2B and @liquidai LFM3 1.7B on instruction following, and it's ~2GB at Q4.
Lab numbers are a good start, but let's see how it holds up on a consumer GPU with a personal agent workflow.
All five go through the same tests on my RTX 3080, with identical llama.cpp settings:
- PA (Personal Agent) Core agent tasks
- Tool calling discipline
- Context retrieval
- Speed and VRAM fit
Which one are you betting on?👇
Is this the newest and best 2B model available?
The company behind PUBG just turned a 30B MoE into a small AI model you can download. 👀
KRAFTON released Raon-1 nano, designed for conversations, instruction following, and tool use in English and Korean.
They pruned and distilled Raon-1 mini preview, converted it into a dense model, then continued training on 3 trillion tokens before further post-training.
A smaller assistant that follows instructions and keeps track of the conversation could be useful without needing a giant model behind every interaction.
KRAFTON’s reported results ...
📋 IFBench → 75.9
🧠 PersonaMem v2, 32K context → 38.8
📱 AndroidControl → 36.0
It leads the selected comparison models on those tests—not every benchmark.
Release stats ...
💾 Downloadable weights
🔓 Apache 2.0
🌐 English + Korean
🧠 3.1B total parameters, described as an effective 1.5B dense model with per-layer embeddings
GGUFs only <4GB
So now we have to test how fast it runs, how much memory does it need, and how reliably does it handle your tools?
ALT https://huggingface.co/KRAFTON/Raon-1-nano https://huggingface.co/KRAFTON/Raon-1-nano-gguf
my favourite thing in the world of agentic coding - i now have the ability to connect clunky systems, build dashboards and collect data.
this would of been me or someone else living in a spreadsheet or paying for an integration service.
next project inhouse netsuite > shopify connector
27
Follow my new journey dsh-personal-assistant
Can we turn the minimal DSH into a personal assistant? There's lots of plugins in the ecosystem already to make it possible, but been building my own plugins to make it happen.
The github repo is my journal of the journey ahead.
github.com/RegeneratusLabs/d…
1
4
185
Liang Wenfeng the wise
🤖 Made with AI
Meanwhile DeepSeek founder Liang Wenfeng:
• “AI is large enough that it might eventually account for 10% of human GDP. You cannot monopolize such a thing. You must share it with others, or you won’t survive. If we tried to monopolize the benefits, we would be discarded by history. This is an objective law, a historical perspective.”
• “Those who take more get beaten by those who take less. You don’t even need to actually be taking more. If your ambition is to take more, you will be beaten by someone whose ambition is to take less.”
• “If your vision is to take more, you’ve already lost; you may face even greater difficulties. That’s just the way the world is.”
• “We must not imagine that we can claim a certain percentage of the world’s GDP for ourselves — the more you cling to that idea, the less likely you are to succeed. … The more restrained you are, the more likely you are to achieve AGI.”
5
128
Josh retweeted
I just created a PlaudNote clone using a @waveshare00's ESP32-S3 based device
I was thinking of buying a device to record some meetings and take notes using voice, a dedicated device to gather some context, but the prices and subscription costs were not worth it for me, so I decided to make my own.
It is simple: you can record voice using the two internal mics, and on the server side, it transcribes the audio file using Whisper, creates a summary using a local LLM—in this demo I am using @OpenBMB MiniCPM-1B—and syncs the result with my Obsidian vault. Everything is processed locally, with no cloud and no internet.
Contributors and makers: I'll be open-sourcing the repo soon.
❤️❤️❤️
@RightSideOfAI Really appreciate everything you do for the MiniCPM community and your mission to make AI beneficial for everyone. We share your belief that small, efficient models have a huge role to play in the future. Excited to see what you'll build next, and what the community creates together! ❤️
2
139
Not everything has to be run by agents
I met a guy who was integrating AI into a freight company - $2500 per month, cycling through Claude subscriptions. He described running 100s of simultaneous agents to run the behind the scenes work. He said the agents were alerting him too much via email so he's delegated agents to monitor the agents.
First red flag - too many subagents can get out of control very quickly. You probably spend more time and more tokens fixing and monitoring.
How much of those agents work could of just be python scripts or autonomous software?
How many of those subagents could be a cheaper model with appropriate guardrails?
Or is there a product on the shelf that handles everything? And the AI driven components are completely unnecessary?
AI isn't always the option.
1
1
72
Josh retweeted
Introducing the MiniCPM-OCR plugin for Hermes Agent.
This is a problem I have experienced myself: LLMs by themselves are not always very good at OCR; sometimes I have needed another model specialized in it to solve complex document parsing.
For example, recently I have had difficulties understanding some physician recipes and other manuscripts.
This is where MiniCPM-OCR comes in. You can use the model and harness you already have to utilize a best-in-class OCR model by @OpenBMB.
@Teknium @NousResearch
Code: github.com/PovedaAqui/minicp…
PR: github.com/NousResearch/herm…
Have you tried the lazy plugin? create your own plug-in within Hermes, on an interactive UI dashboard, no technical skills required, create it from a cron job, or an existing skill, upload it straight to your GitHub, and submit a PR to the Hermes team. It's that simple,
my @n Hermes hackathon contribution.
Open to contributions to make this even better and more polished an ideal outcome for this would be a Hermes chat scanner that identifies potential missing plug-ins based on your previous chat history.
github.com/RegeneratusLabs/l…
lazy-plugin live action.
idea > plugin > tested > PR > github repo
this plugin creation started with an existing cron job i've had set up, my GPUs under $600 cron job I run. but you can convert your existing skills or just describe it in natural language.
session is spawned and the plugin is built right in your hermes, you can steer the conversation and answer clarifying questions in the chat session.
then publish it straight to the hermes PR
my hermes hackathon submission
@witcheer you might like this one tooo!
49
My second @NousResearch Hermes hackathon submission
stash, see something on X, YouTube, Reddit, or the web, including images, copy the link and paste it into stash, and stash it for a future Hermes.
I built this because I was finding I would find information and I would bookmark it on a browser, or I would start a session with Hermes and I'd never come back to it. Or there was too much information in too many different places. Now it all just lives in your Hermes UI and you can come back to it and start a session with Hermes based on the stashed item.
Open to contributions to make it more polished.
github.com/RegeneratusLabs/h…
stash hermes plugin.
it's that simple. keep your hermes ideas such as @HermesWatcher posts in your hermes desktop.
a hermes hackathon submission
github.com/RegeneratusLabs/h…
5
17
873
DSH is pretty smooth with PWA. I run everything on the web ui so it's completely seamless between devices.
49
Josh retweeted
Introducing ERNIE Evolve — our automated research system for recursive self-improvement (RSI).
Across the three metrics reported for NanoChat AutoResearch and NanoGPT Speedrun, ERNIE Evolve surpassed the best publicly verified results available at the time of evaluation — and it already helped build ERNIE 5.5 Flash.
My main concern at the moment with DSH and the plug-in ecosystem, who's vetting it for security vulnerabilities. There's a lot of plug-ins and a lot of them I wouldn't touch with a 10-foot pole, especially when it comes to telemetry and privacy or connecting to unknown servers. Just remember to always check what you install or build your own from scratch.
3
135
