Hot take: Closed AI models are a bit overhyped and overpriced. A quick comparison with recently launched open weights models ↓ GLM 5.2: – $1.40 in / $4.40 out per 1M tokens – 1M context – SWE-Bench Pro: 62.1% (beats GPT-5.5 by 3.5 pts) – 1/6th the price MiniMax M3: – $0.60 in / $2.40 out per 1M tokens – 1M context – SWE-Bench Pro: 59% (Opus 4.8: 69.2%, GPT-5.5: 58.6%) – 1/8th the price Kimi K2.6: – $0.95 in / $4.00 out per 1M tokens – 256K context – SWE-Bench Pro: 58.6% (ties GPT-5.5) –…