GLM-5.3
22nd on WebDev at 1622, MIT. 20th on Terminal-Bench at 41.8% ± 3.2%, cost $2.7k. Weights can be self-hosted. 7th on DeepSWE v1.1 Best (max) at 69.0%.
Zhipu · MIT · $1.40 / $4.40 per million tokens · Fetched 2026-10-10
Benchmark scores
Boards older than 90 days are not counted as recommendationsBenchmarkSettingScoreRankCostAs of
Use it or not
Why
- MIT. The open-weight pick on Terminal-Bench.
When not
- 41.8% on Terminal-Bench, well behind the proprietary front.