GLM-5.3 Flash vs Gemini 3.8 Flash
Too close to call. On the work we tested, GLM-5.3 Flash and Gemini 3.8 Flash can't be told apart. GLM-5.3 Flash is the cheaper one, by 6.6×.
GrindBench-PubSec
BGLM-5.3 FlashZ.ai (Zhipu)61% 56–67%0.53¢ a task · 300 tasksCGemini 3.8 FlashGoogle52% 47–58%2.1¢ 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 | GLM-5.3 Flash | Gemini 3.8 Flash |
|---|---|---|
| ATIP redaction | 20% | 18% |
| Procurement evaluation | 78% | 81% |
| Performance analysis | 57% | 33% |
| Bilingual equivalence | 92% | 97% |
GrindBench-Banking
CGLM-5.3 FlashZ.ai (Zhipu)47% 35–60%1.4¢ a task · 60 tasksCGemini 3.8 FlashGoogle48% 35–60%$0.11 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 | GLM-5.3 Flash | Gemini 3.8 Flash |
|---|---|---|
| Lending & mortgages | 50% | 50% |
| AML & KYC | 50% | 30% |
| Deposits & payments | 50% | 40% |
| Complaints & conduct | 70% | 60% |
| Credit risk | 20% | 60% |
| Capital & liquidity | 40% | 50% |
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.