It's an okay model. My biggest issue using GLM 5.1 in OpenCode is that it loses coherency over longer contexts. When you crest 128k tokens, there's a high chance that the model will start spouting gibberish until you compact the history.
For short-term bugfixing and tweaks though, it does about what I'd expect from Sonnet for a pretty low price.
Unsloth quantizations are available on release as well. [0] The IQ4_XS is a massive 361 GB with the 754B parameters. This is definitely a model your average local LLM enthusiast is not going to be able to run even with high end hardware.
I am on their "Coding Lite" plan, which I got a lot of use out of for a few months, but it has been seriously gimped now. Obvious quantization issues, going in circles, flipping from X to !X, injecting chinese characters. It is useless now for any serious coding work.
I have their most expensive plan and it's on-par and sometimes better than Claude although you have to keep context short. That being said, the quota is no longer generous. It's still priced below Claude but not by that much. (compared to a few months ago where your money gets you x10 in tokens)
To be honest I am a bit sad as, glm5.1 is producing mich better typescript than opus or codex imo, but no matter what it does sometimes go into shizo mode at some point over longer contexts. Not always tho I have had multiple session go over 200k and be fine.
I am already subscribed to their GLM Coding Pro monthly plan and working with GLM 5.1 coupled with Open Code is such a pleasure! I will cancel my Cursor subscription.
Comments here seem to be talking like they've used this model for longer than a few hours -- is this true, or are y'all just sharing your initial thoughts?
GLM-5.0 is the real deal as far as open source models go. In our internal benchmarks it consistently outperforms other open source models, and was on par with things like GPT-5.2. Note that we don't use it for coding - we use it for more fuzzy tasks.
It really bothers me that people refer to open weight models as being open source. They fundamentally aren't and are more akin to freeware than anything else.
The focus on the speed of the agent generated code as a measure of model quality is unusual and interesting. I've been focusing on intentionally benchmaxxing agentic projects (e.g. "create benchmarks, get a baseline, then make the benchmarks 1.4x faster or better without cheating the benchmarks or causing any regression in output quality") and Opus 4.6 does it very well: in Rust, it can find enough low-level optimizations to make already-fast Rust code up to 6x faster while still passing all tests.
It's a fun way to quantify the real-world performance between models that's more practical and actionable.
Feeling very much the same. Attempting to use it through Claude Code as a model it just completely lost all context on what it was doing after a few months and kept short circuiting even with the most helpful prompts I could give, outside of just writing out the answer myself. I really do not get the praise for this model.
Being "better than Opus 4.6" is not really something a benchmark will tell you. It's much more a consensus of users liking the flavor of an answer, rather than fueling x% correct on a benchmark.
I'm crossing my fingers they release a flash version of this. GLM 4.7 Flash is the main model I use locally for agentic coding work, it's pretty incredible. Didn't find anything in the release about it - but hoping it's on the horizon.
I can’t wait to try it. I set up a new system this morning with OpenClaw and GLM-5, and I like GLM-5 as the backend for Claude Code. Excellent results.
A bit off-topic, but for some reason, even though I don't use LLMs for my job or for my hobbies, or in daily life frequently (and when I do, it's mostly some kind of "rubber duck brainstorm"), when I see open-weight releases like this one or the recent Gemma 4 (which is very good for local models); the first time was with DeepSeek-R1 (this one, despite being blamed for "censorship", was heavily censored only via DeepSeek API, the local model - full-weight 685B, not the distilled ones - was pretty much unhinged regarding censorship on any topic)... there's always one song coming to mind and I simply can't get rid of it no matter how hard I try.
"I am the storm that is approaching, provoking..." : )
Every single day, three things are becoming more and more clear:
(1) OpenAI & Anthropic are absolutely cooked; it's obvious they have no moat
(2) Local/private inference is the future of AI
(3) There's *still* no killer product yet (so get to work!)
45 comments
[ 2.8 ms ] story [ 56.9 ms ] thread[[you guys, please don't post like this to HN - it will just irritate the community and get you flamed]]
For short-term bugfixing and tweaks though, it does about what I'd expect from Sonnet for a pretty low price.
[0] https://huggingface.co/unsloth/GLM-5.1-GGUF
We all know that building a spec-compliant browser alone is a herculean task.
It's a fun way to quantify the real-world performance between models that's more practical and actionable.
Being "better than Opus 4.6" is not really something a benchmark will tell you. It's much more a consensus of users liking the flavor of an answer, rather than fueling x% correct on a benchmark.
Excited to test this.
"I am the storm that is approaching, provoking..." : )
Everyone else isn't that far behind and they aren't all gonna just wall off their new model.
A reason that Anthropic will eventually give is 'the competition can do what Glasswing can do so what's the point limiting it'.