Z.ai releases GLM-5.1, a 754B-parameter Mixture-of-Experts model that it says outperforms GPT-5.4 and Opus 4.6 on SWE-Bench Pro, available under an MIT license
Is China picking back up the open source AI baton? — Z.ai, also known as Zhupai AI, a Chinese AI startup best known for its powerful …
VentureBeatCarl Franzen
Context & Ripple Effects
Z.ai has moved through lower-cost open-source positioning with GLM-4.5 to a flagship GLM-5 claim of leading open-source performance in reasoning, coding, and agentic work. GLM-5.1 extends that progression with a much larger Mixture-of-Experts release and an MIT license.
The significance is not the benchmark claim alone, which remains Z.ai’s assertion, but the combination of frontier-oriented coding evaluation and permissive access. It puts an openly licensed alternative alongside proprietary systems whose product differentiation also rests on capabilities such as Claude’s expanded long-context access.
First-order effects
Developers and enterprises can inspect, modify, and deploy GLM-5.1 under MIT terms rather than relying solely on hosted proprietary models, subject to their own infrastructure capacity.
Open-model vendors and proprietary API providers face greater pressure to distinguish their offers through reliability, tooling, long-context workflows, support, and deployment convenience—not benchmark performance alone.
A permissively licensed model with claimed high coding performance can strengthen demand for the compute, hosting, integration, and evaluation layers needed to run models outside a single provider’s API.
Third-order effects
If comparable open-weight models continue to close the gap on proprietary systems, model access may shift from a product controlled mainly through APIs toward a more distributed software and infrastructure market.
The competitive boundary would increasingly be set by deployment ecosystems and access rules, linking model releases to broader model-access geopolitics rather than parameter scale alone.
The trend: This is one data point in the industrialization of AI: frontier-oriented capabilities are being paired with more permissive model access, moving competition downstream into deployment and distribution.
we open-sourced glm-5.1 agents could do about 20 steps by the end of last year. glm-5.1 can do 1,700 rn. autonomous work time may be the most important curve after scaling laws. glm-5.1 will be the first point on that curve that the open-source community can verify with their own
GLM-5.1 is here! Try it on OpenClaw🦞🦞🦞 ollama launch openclaw —model glm-5.1:cloud Claude Code ollama launch claude —model glm-5.1:cloud Chat with the model ollama run glm-5.1:cloud
Building a Linux Desktop from Scratch Using a self-review loop, GLM-5.1 spent 8 hours autonomously refining features, styling, and interactions to build a functional desktop environment. [video]
Wow, GLM-5.1 beat Opus 4.6, GPT-5.4, and Gemini 3.1 Pro on SWE-Bench Pro (58.4 vs 57.3 / 57.7 / 54.2) as an open-weight MIT-licensed model! The “open-source AI vs closed-source AI” gap is still ~6 months. [image]
Introducing GLM-5.1: The Next Level of Open Source - Top-Tier Performance: #1 in open source and #3 globally across SWE-Bench Pro, Terminal-Bench, and NL2Repo. - Built for Long-Horizon Tasks: Runs autonomously for 8 hours, refining strategies through thousands of iterations. [ima…
Vector-DB-Bench: 6x Performance Boost In high-performance database optimization, GLM-5.1 reached 21.5k QPS over 600+ iterations and 6,000+ tool calls. This is 6x the performance of a standard 50-turn session. [image]
A new open model has entered the Arena! GLM-5.1 by @Zai_org is now ready for your prompts in the Text and Code Arena. Come vote and let's see how it stacks up! [image]
Another big release: GLM-5.1! China is on fire! significant increase in evals compared to GLM-5.0 tl;dr GLM-5.1 is the new open-source agentic coding model that significantly outperforms its predecessor by sustaining long-horizon problem-solving over hundreds of iterations, [imag…
This is crazy! https://z.ai/ has caught up with the SOTA models in coding with its GLM-5.1 open-weight model. This is a very big deal given that billions of $ are now spent on coding tokens! Congratulations to @Zai_org team, major achievement! [image]
Zhipu raised the cost of access to its most advanced AI model by at least 8%, joining other leading Chinese AI players in trying to profit off years of research and computing investments https://www.bloomberg.com/...
Most models still break mid-task not because they're not smart enough but because they can't stay in the loop 8-hour runs start to change that this is how agents stop breaking.
GLM-5.1 by @Zai_org just launched in the Text Arena, and is now the #1 open model. It outperforms the next best open model, its predecessor, GLM-5, by +11 points and +15 over Kimi K2.5 Thinking. It shows strength in: - #1 open model in Longer Query (#4 overall) - #1 open model [i…
INCREDIBLE GLM-5.1 weights are now opensource > i've had early access to the weights for the past few days > and yeah... this one matters a lot benchmarks? > SWE-Bench Pro: 58.4 > beats Opus 4.6 (57.3) > beats GPT-5.4 (57.7) > beats Gemini 3.1 Pro (54.2) let that sink in [image]
Holy moly, we thought Tuesday would be dull, but GLM-5.1 is out and it's freaking open source too 😲 What will it take to run something like this locally? A fortune? [image]
🎉 Day-0 support for GLM-5.1 in vLLM! Congrats to @Zai_org on this next-gen flagship model built for agentic engineering, with stronger coding and sustained long-horizon task performance. Get started 👇 📖 Recipe: https://docs.vllm.ai/... [image]
Venice delivers in <5m Available for Free to Pro users and DIEM holders in API (this model is killer for agents... first one that feels anecdotally comparable to opus imho)
damn new GLM-5.1 model crushes anthropic and openai at agentic coding and 100% open source, how does china keep getting away with this shit? - beat opus 4.6 by 6X on an open-ended coding problem - long-task horizon: 600 turns at a time + 1000s of tool calls. usually ai's just [im…