Dylan PatelFounder and Chief Analyst, SemiAnalysis · The semiconductor-AI stack

GLM 5.3’s biggest leap came from post-training alone.

17 Oct 2026 · 2 min

Watch on YouTube

1:50

In this brief multi‑speaker discussion, Speaker A reflects on the release of GLM 5.3, emphasizing that its >10% benchmark improvements across agentic reasoning, general reinforcement learning, and cybersecurity stem entirely from pulse (post‑training) applied to previous GLM models. He highlights the method’s GPU efficiency and its appeal to companies with limited budgets but proprietary data, suggesting they can achieve superior performance without full model retraining. Speaker B concurs, noting the dramatic “night‑and‑day” boost in cybersecurity capabilities from GLM 5.2 to 5.3, again attributing the gains to post‑training. Both speakers present the development as a wake‑up call for the industry.

Chapters

Actually, I want to say what I feel about GLM 5.3 when it first came out. It's completely based on pulse training with the previous models. I feel like this is like a wake-up call for me. So if you see the benchmark results of GLM 5.3, it's gigantic results on the benchmark, was improved more than 10% by just doing pulse training. And it's more GPU efficient if you're doing pulse training than pretrained model from scratch. And I think that's a wake-up call for a lot of other companies as well, that they don't have enough budget to train a model from scratch, but they also want to improve their company's workflow for some specific area. And they happen to have that proprietary and private data, so they can just do post-training themselves and achieve a better result. Than some of the closed-source models. If I remember correctly, that GLM 5.3 was focusing on agentic reasoning and general RL, and we see all the 3 areas has improved quite a lot by looking through the benchmark results. So that's like the day that I saw the Twitter where GLM 5.3 comes out and all the benchmark improvement, I was like, no way. Just done this by post-training. That's quite, quite amazing to me.
I totally agree with Weinyao on how much juice there's left to squeeze by just post-training. 5.3's post-training effort is definitely very impressive. And also cybersecurity. 5.2, in our article, I think we also compared GLM 5.3's cyber capabilities with 5.2. It's like very notable, like a night and day capability difference. And that's all just post-training.