Abel🪶
497 posts
Abel🪶
@thunderbloo
Abel Liu, MD. Long-term resident of hospitals, codebases, and random late-night experiments.
Bergabung Ocak 2014
363 Mengikuti50 Pengikut
This is what I was talking about.
DeepSeek is extremely smart.
Tim@TimGMath
The latest versions of ArXivMath and BrokenArXiv have been released! Impressive Performance of Fable 5, which takes the top spot on ArXivMath. On BrokenArXiv, GPT 5.5 continues to be in the lead.
English
Still don't understand why China doesn't get to host the world cup
People love football, they have infrastructure, it's cheap, VISA free, good time zone
Would easily be the most popular edition (in terms of visitors)
Is FIFA boycotting? Why?
English
I think DeepSeek feels smarter overall, but sometimes it seems to overthink surprisingly simple things and get stuck going in circles.
MIMO feels almost like the opposite. On certain tasks, it suddenly becomes extremely smooth and effective, as if it was particularly well-tuned for them. But overall, I still prefer DeepSeek.
English
@thunderbloo How was the performance? I only tried OpenCode Go so would love to know your thoughts on DeepSeek API vs Mimo token plan
English
迭代速度太快了。如此看来，OpenClaw 有一个衰败的原因之一是创始人单打独斗很快被挖走了。开源社区的方式太慢了，质量也一般，不如这种公司创业团队作战快呀，开源一败涂地。
Nous Research@NousResearch
Hermes Agent now has Automation Blueprints, turning cron jobs into clickable, fillable, conversational workflows.
中文
Nex-N2 Pro feels very GPT-coded.
The frontend outputs look extremely similar to what I usually see from OpenAI models lol. The structure, UI polish, component style, even the way it approaches app layouts.
Looks like the GPT-style outputs were heavily used in post-training. Full Breakdown: youtu.be/n4XMisTcL0k
Either way, the model is free and solid… just interesting. 👀
YouTube
ModelScope@ModelScope2022
Nex-N2 is now open source！An agentic model series from Nex AGI built for coding, tool use, deep research, and long-horizon workflows. 🧠🔎 🛠️ modelscope.ai/models/nex-agi… ⚙️ modelscope.ai/models/nex-agi… ● Models: Nex-N2-Pro 397B total, 17B active; Nex-N2-mini 35B total, 3B active ● Agentic Thinking: adaptive reasoning depth + coherent reasoning across coding, search, tool calling, and execution ● Efficiency: Nex-N2-mini saves roughly 20% overall token cost vs forced thinking while matching or slightly exceeding task performance ● Open-model lead: 75.3 on Terminal-Bench 2.1, 80.8 on SWE-Bench Verified, 83.7 on BrowseComp, and 1585 on GDPval among listed open baselines ● Deployment: customized SGLang fork, reasoning parser, tool-call parser, Docker image ● License: Apache 2.0
English
What is this again, OMG
Nex@NexEcosystem
📢 Nex-N2 is here! A family of agentic models that doesn't just think, it acts! Coding, search, tool use. All fused into a single agentic reasoning loop. - Adaptive Thinking, auto-scales reasoning depth per step. Saves ~20% tokens, zero performance loss. - Coherent Thinking, one thinking paradigm across search, coding, and tool use. No more fragile mode-switching. 🏆 Result: Tier-1 open-source performance on SWE-bench, Terminal-Bench, GDPval, and more, tracking GPT-5.5 and Opus 4.7. 🎉 Open-weight. Try it now. 🔗 nex-agi.com 📦 huggingface.co/nex-agi/Nex-N2… modelscope.cn/models/nex-agi… github.com/nex-agi/Nex-N2
English
@freeguardteam issue 看描述只要长期运行都有泄露的，issue 已经有人发过，看样子是有几个月的老问题了。
AI 查了一下意思是代理本身没问题，是 GUI 轮询流量累积不释放的 bug。
中文
@thunderbloo 可以先把 TUN 和客户端进程拆开看：记录 Clash Verge 版本和运行时长，关 TUN 跑同样订阅做对照；开 TUN 后看进程 RSS、连接数和 DNS/fake-ip/LAN 排除。如果只有 TUN 长时间运行持续涨，带版本、日志和截图发 issue 会比换节点更有用。
中文
@teortaxesTex Change the reset from 5 hours to 1 hour.
Problem solved.
English
why not? I'm not asking them to *train it* in a day.
ρ:ɡeσn@pigeon__s
@teortaxesTex its literally been one day bro did you expect gpt-5.6 to release 5 mins after????
English
用了一下午，是比较水桶的模型。
写 demo 几乎无短板，大一点的 codebase 聊天体验比较好，但是给出的建议相对一般。。。。
压力给到 奥特曼这边
海拉鲁编程客@hylarucoder
fable 5 肯定比 opus 4.8 好很多。 但比4.6好多少就不清楚了
中文