GLM-4.7

By price, GLM-4.7 ranks 24th cheapest of 68 on input and 31st on output. Its quality score of 87 ranks 33rd of 68, and on value (quality score ÷ output price) it ranks 30th of 68. Its blended rate sits 73% below the catalog average.

In real billing terms: a support chatbot handling 50,000 turns a month (2,000 input / 500 output tokens each) costs about $84 per month on GLM-4.7, while a RAG workload of 20,000 queries (8,000 input / 800 output tokens each) runs about $92. Because output is priced 4.4× input here, long-generation tasks scale cost fastest — cap output length before you optimise anything else.

A 205K context window holds roughly 273 pages of A4 text (at ~750 tokens per page), ranking 42nd of 68 here. That covers typical RAG and multi-turn chat, though book-length inputs still need retrieval and chunking. Output runs at about 130 tok/s (21st fastest of 68), fast enough for real-time interaction and high-throughput batches.

We track 2 Zhipu models, and GLM-4.7's blended rate sits 49% below that provider's average. If your task tolerates a little less headroom, Hunyuan HY3 (Tencent) blends about 86% cheaper at a quality score of 87, and is worth benchmarking first. Going the other way, qwen3.5 scores 91 on quality for about 85% less.

The closest model in Zhipu's line-up is GLM-5.2 ($1.40 input / $4.40 output, quality 93). Against it, GLM-4.7 scores 6 points lower on quality, runs 30 tok/s faster (about 30%), has a 795K smaller context window, costs 192% less blended. When two models sit this close on price, the deciding factor is usually the quality ceiling — run both against your real prompts before picking.