@tphuang it overthinks, although gets it right at the end, but man does it over think.
English
Nour Eddine Hamaidi
2.1K posts
Qwen 3.6 Plus looks to be quite the upgrade on coding & agentic stuff vs Qwen 3.5 open src models & that has brought in a lot of traffic.
According to coding arena, it's on same level as GPT-5.4 Codex & Gemini-3.1 pro & better than GLM. Quite the upgrade over 3.5
OpenRouter@OpenRouter
Qwen 3.6 Plus from @Alibaba_Qwen is officially the first model on OpenRouter to break 1 Trillion tokens processed in a single day! At ~1,400,000,000,000 tokens, it’s the strongest full day performance of any new model dropped this year. Congrats to the Qwen team!
English
Theory, anthropic throttles claude to 1tok/s when they need more revenue so they can invalidate all 5min and 1hr ttl caches lol
English
@Rasmic I'd say glm-5-turbo on Hermes Agent. It's not even close.
English
You can try GLM-5-Turbo and GLM-5V-Turbo.
Maybe they are not quite Claude Opus 4.6 level yet, but they are absolutely strong contenders — and definitely worth a shot.
Boris Cherny@bcherny
Starting tomorrow at 12pm PT, Claude subscriptions will no longer cover usage on third-party tools like OpenClaw. You can still use these tools with your Claude login via extra usage bundles (now available at a discount), or with a Claude API key.
English
@TheMarshmalon @honchodotdev Super interesting, never heard of it, but worth checking.
English
Happy to help anyone out there with questions about @honchodotdev
My DMs are open
Teknium (e/λ)@Teknium
Hermes Agent now supports @plastic_lab's Honcho, @mem0ai, @openvikingai, @Vectorizeio's Hindsight, @retaindb, and @ByteroverDev memory systems! Try them now with `hermes update` then `hermes memory setup` We have rehauled our memory system to be much more maintainable and pluggable, so anyone can make their own memory system to build on top of Hermes easily and cleanly with a special class of plugin! Which memory system is your favorite?
English
We also added in-line diffs to the CLI TUI!
Many coders desired the ability to see the changes the agent was making when it was making it, to better help them know what's going on.
Now you can! It is on by default but able to be disabled in the config!
English
Deeper dive into some of the updates in v0.7
Memory:
We have begun transitioning each of the systems in Hermes Agent to work through defined interfaces so that the core code is more maintainable, and more providers for everything can be supported.
We started with memory:
Now there is a plugin type that is for memory providers, pulled Honcho out of the core codebase, made it a plugin, and added 6 more systems that were waiting as PRs as new options.
Now you can access your favorite much easier, and if your favorite isn't built in, it is easy to create a new plugin that does!
Nous Research@NousResearch
Hermes Agent v0.7.0 is out now. Our headline update: Memory is now an extensible plugin system. Swap in any backend, or build your own. Built-in memory works out of the box; six third-party providers are ready to go. Pick one with 'hermes memory setup'. Full changelog below ↓
English
@kevin_kamal8 @mark_fietje The solution is on the post, just use Hermes Agent and give it the context of my post and it shall know what to do.
English
@mark_fietje @NOOROU Hello
Have any of you come to a solution?
A few days ago I have been trying to use AutoClaw but unfortunately it asks me for a Chinese phone number
English
The Hermes Agent x GLM-5-Turbo challenge:
Hermes Agent just reversed engeenerd AutoClaw. It's a clean Electron app with a local gateway running on port 18789 that proxies calls through their API.
The only catch for non-Chinese users is the login screen asking for a Chinese phone number. There's no API key option, no alternative auth, and no way to skip it.
Hermes Agent handled this in about 5 minutes of tinkering
English
@Teknium in a month from now you are going to see a series of big days that make this look like a joke. Hermes ftw.
English
Our biggest day EVER with Hermes Agent, we're now #5 biggest AI App on OpenRouter metrics!
What do you want to see in the next update?
English
GLM 5V Turbo ranks #14 on BridgeBench Hallucination.
71.4 score.
34.2% fabrication rate.
Fast? Top 5 on SpeedBench. Honest? Not even close.
A third of the time it's confidently making things up about your code.
Same fabrication rate as Claude Haiku 4.5.
Z.ai fixed the speed problem from GLM 5.1.
They didn't fix the trust problem.
Speed without accuracy is just hallucinating faster.
bridgebench.ai
English
Introducing GLM-5V-Turbo: Vision Coding Model
- Native Multimodal Coding: Natively understands multimodal inputs including images, videos, design drafts, and document layouts.
- Balanced Visual and Programming Capabilities: Achieves leading performance across core benchmarks for multimodal coding, tool use, and GUI Agents.
- Deep Adaptation for Claude Code and Claw Scenarios: Works in deep synergy with Agents like Claude Code and OpenClaw.
Try it now: chat.z.ai
API: docs.z.ai/guides/vlm/glm…
Coding Plan trial applications: docs.google.com/forms/d/e/1FAI…
English
Due to the leakage of Claude Code, the following projects were generated:
github.com/instructkr/cla…, Star 94.3K, Fork 88.5k
github.com/claude-code-be…, Star 5.7K, Fork 7.2k
github.com/paoloanzn/free…, Star 2.4K, Fork 930k
github.com/NanmiCoder/cla…, Star 1.8K, Fork 2.4k
github.com/oboard/claude-…, Star 1.7K, Fork 2.5k
English
I feel like Anthropic did this on Purpose to get free army of developers with ideas and millions of tokesn on refactors. a super WIN-WIN. There's a Rust version already in development for FREE to them.
English
@HARRY_1G I think this might be their wierd way of opensouring it. Still a conspiracy theory though.
English
@NOOROU If they send DMCAs again they're gonna lose the little respect they have left I fear
English
Finally. Claude Codex.
Now I can use codex with my subscription as well as claude sub in claude code.
Another open source W.
English
@Teknium Hermes is catching up at an unimaginable speed. We really need extra hands for PR reviews.
English
Will be reviewing this PR today!
0xByte@0xbyt4
built computer use into hermes agent tell it what to do from your phone → it controls your mac no sandbox real desktop, real apps, real time a complex example : diagram on freeform @Teknium @NousResearch @claudeai for more : github.com/NousResearch/h…
English
Does @NousResearch welcome and actually look ay external contributions to Hermes?
My biggest gripe right now with OpenClaw is the inability to get any bug fixes looked at/reviewed/merged.
I'm maintaining a local fork at this point because I gave up on getting OTEL fixed in main
English
@bridgemindai Does the distance from their servers have anything to do with this? I mean in my case it is isn't that blizzard.
English
GLM 5.1 has one of the best UI skills of any coding model I've tested.
It still isn't ready.
Dropped $80 on the GLM Max Coding Plan.
Put it through real production workflows in BridgeSpace.
The UI output is better than GPT 5.4.
Better than Claude Opus 4.6 on design tasks.
That part shocked me.
But 30 seconds to read 5 lines of a file.
Constant 529 errors. Tiny context window.
Slowest model on BridgeBench SpeedBench.
Beautiful code you can't ship fast enough is still code you can't ship.
Full review and exclusive BridgeBench results below.
English