| Model | Score | Status | Released | Price in/out | Context | |
|---|---|---|---|---|---|---|
| GLM-5.2 open-weights frontier coding agents | 8.8 | GA | 2026-06-15 | $1.40 / $4.40 | 1M | Review → |
GLM-5.2 is the moment open weights stopped trailing: Z.ai's flagship 753B-parameter MoE (40B active) beats GPT-5.5 on multiple long-horizon coding benchmarks — SWE-bench Pro 62.1 vs 58.6 — at roughly one sixth the price, under a plain MIT license with a 1M-token context window. Released 2026-06-16, it took the #1 open-weights slot on the Artificial Analysis Intelligence Index (51) within a week, and its selectable reasoning effort plus visible thinking make it the most credible self-hostable alternative to the closed frontier as of mid-2026.
Full review →