zai-glm-4.6 vs zai-glm-4.7

Capability for the money

Both are priced within 15% of each other, so capability decides it: zai-glm-4.7 leads by 31.2 points on following instructions (IFBench).

1. Following instructions

   extraction, classification, sticking to a format

   IFBench
   zai-glm-4.636.7%

   zai-glm-4.767.9%

   zai-glm-4.7 +31.2

2. Coding

   writing and fixing code

   LiveCodeBench
   zai-glm-4.656.1%

   zai-glm-4.789.4%

   zai-glm-4.7 +33.3

3. Tool use

   multi-step agent work against real tools

   τ²-bench
   zai-glm-4.676.9%

   zai-glm-4.795.9%

   zai-glm-4.7 +19.0

4. Hard reasoning

   graduate-level science and multi-step analysis

   GPQA Diamond
   zai-glm-4.663.2%

   zai-glm-4.785.9%

   zai-glm-4.7 +22.7

Averaged over the benchmarks both models sat, zai-glm-4.7 returns 35.7 points of pass rate per dollar of blended price against 24.5 — the ratio that decides whether the dearer model earns its rate.

Each figure is a pass rate: the share of that test set the model answered correctly, measured by a third party rather than self-reported by the lab. Rows are different tests of different difficulty, so compare a model against the other model, not one row against the next. And a benchmark is not your workload — it narrows the shortlist, it does not pick for you.