Mercury 3 vs GLM-5.3 Flash
Too close to call. On the work we tested, Mercury 3 and GLM-5.3 Flash can't be told apart. Mercury 3 is the cheaper one, slightly.
GrindBench-PubSec
BMercury 3Inception61% 56–67%0.53¢ a task · 300 tasksBGLM-5.3 FlashZ.ai (Zhipu)61% 56–67%0.53¢ a task · 300 tasks
Too close to call here: their ranges overlap, so on another set of tasks the order could flip.
| Kind of work | Mercury 3 | GLM-5.3 Flash |
|---|---|---|
| ATIP redaction | 11% | 20% |
| Procurement evaluation | 86% | 78% |
| Performance analysis | 54% | 57% |
| Bilingual equivalence | 97% | 92% |
GrindBench-Banking
CMercury 3Inception52% 38–63%1.4¢ a task · 60 tasksCGLM-5.3 FlashZ.ai (Zhipu)47% 35–60%1.4¢ a task · 60 tasks
Too close to call here: their ranges overlap, so on another set of tasks the order could flip.
| Kind of work | Mercury 3 | GLM-5.3 Flash |
|---|---|---|
| Lending & mortgages | 50% | 50% |
| AML & KYC | 40% | 50% |
| Deposits & payments | 40% | 50% |
| Complaints & conduct | 70% | 70% |
| Credit risk | 60% | 20% |
| Capital & liquidity | 50% | 40% |
Each kind of work here rests on 10 tasks per model, so a single row can easily be 20 points off. A lead of 10 points or more is in bold; read it as a hint, not a result.