List of AI News about GLM 5.2
| Time | Details |
|---|---|
|
2026-06-29 16:59 |
DeepSeek V4 Pro tops SWE-bench, real gains need harness
According to @_avichawla, DeepSeek V4 Pro leads SWE-bench Verified, but real coding performance depends on the harness, not just leaderboard scores. |
|
2026-06-24 18:55 |
Gemini 3 Pro tops ARC‑AGI‑2 with 31%
According to Ethan Mollick, Gemini 3 Pro first hit 31% on ARC-AGI-2 in Nov 2025, with an 8–12 month lead over open models like GLM 5.2 at 22.8%. |