OPERATOR READVERDICT · JUN 28, 2026 · 7 MIN
GLM-5.2 Cleared the Six Hard Tasks I Use to Vet Any Cheap Model
A new open-weights model matched my flagship on objective hard tasks. The battery I run before trusting any cheap model in production did not change.
I've been trying it out via OpenRouter, which has it from 9 different providers, almost all of which are charging $1.40/million for input and $4.40/million for output.
6/6
objective hard tasks where the cheap model matched my flagship
GLM-5.2 vs CLAUDE OPUS 4.8
Named stake
What AutoKaam Thinks
- A leaderboard rank is a reason to test a model, not a reason to trust it. I keep a fixed battery of objective hard tasks with execution-checked answers and run it on every cheap or open release bef…
- On my June battery, GLM-5.2 matched Claude Opus 4.8 on all six tasks, from discrete optimization with a proof to a money-rounding fix.
- The flagship still wins where the battery cannot reach, on long autonomous runs, on judgment and voice, and on depth of native tool use. Those decide where each tier runs, not the per-task score.
- Pricing is the easy part. GLM-5.2 lists near $1.40 input and $4.40 output per million tokens versus $5 and $25 for Opus 4.8. The hard part is proving parity on your own work first.
