Zhipu AI drops surprise GLM-5.1 update to challenge Claude 4.6 in agentic coding
Zhipu AI just surprise-dropped GLM-5.1. It rivals Claude 4.6 in coding at 1/7th the price. Here’s how to test it today.
Zhipu AI just dropped a surprise release today with GLM-5.1. The new model arrived completely unannounced and is immediately available to all developers on the GLM Coding Plan.
This rapid iteration comes just six weeks after the massive GLM-5 launch, and two weeks after the release of GLM-5 Turbo. Z.ai has shifted the model’s focus away from general text generation and directly toward reliable, complex agentic engineering.
- The performance jump: Zhipu AI claims a score of 45.3 on internal coding benchmarks, up nearly 10 points from the February release.
- Closing the gap: The new score puts it in direct competition with Anthropic’s Claude Opus 4.6, which scores 47.9 on similar evaluations.
- The architecture: GLM-5.1 builds on a 744-billion parameter MoE foundation trained on roughly 28.5 trillion tokens.
- Instant integration: Developers can plug it straight into tools like Claude Code or OpenClaw by manually changing their target model in the config file to
glm-5.1.
The community reaction is heavily focused on how this rapid update impacts the current pricing monopoly held by Western frontier models. Z.ai is moving aggressively to capture the developer market with significantly lower API costs.
- The cost advantage: Developers are already testing it as a replacement for premium models, citing that it costs roughly one-seventh the price of Anthropic’s platform for similar coding workflows.
- Open-source plans: The company’s global lead confirmed that the weights for GLM-5.1 will be open-sourced, though exact timing is still pending.
- Rollout bugs: Because it was a surprise drop, early adopters are reporting temporary “no permission” glitches and higher-than-expected quota consumption.
Bottom line: If you are already on the GLM Coding Plan, it is worth updating your config file to test the new capabilities on your own workflows.
If you need on-demand GPUs for training, fine-tuning, inference, or running open-source models, give RunPod a try.
- Available hardware: H100, H200, A100, L40S, RTX 4090, RTX 5090, and 30+ more
- Cost: significantly cheaper than AWS or GCP, billed per second, no contracts
- Setup: spins up in under a minute, 30+ regions worldwide
Get the core business tech news delivered straight to your inbox. We track AI, automation, SaaS, and cybersecurity so you don't have to.
Just read what you want, and be done with it.
