EvoQwen2.5 VL Retriever 3B v1 vs EvoQwen2.5 VL Retriever 7B v1
No benchmarks in this slice.
Specs
| Spec | EvoQwen2.5 VL Retriever 3B v1 | EvoQwen2.5 VL Retriever 7B v1 |
|---|---|---|
| Context | 128K | 128K |
| Weights | Open | Open |
| Released | Nov 2025 | Nov 2025 |
| Dimensions | 2,048 | 3,584 |
| Max input | 128K tokens | 128K tokens |
| Languages | en | en |
| Memory (4-bit) | 1.8 GB | 4.2 GB |
| Licence | Apache-2.0 | Apache-2.0 |
| Commercial use | Allowed | Allowed |
What breaks when you switch
EvoQwen2.5 VL Retriever 3B v1 → EvoQwen2.5 VL Retriever 7B v1
Re-embed your whole corpus›
Vectors from different models can't be mixed; dimensions change from 2,048 to 3,584.
Memory (4-bit): 1.8 GB → 4.2 GB
EvoQwen2.5 VL Retriever 7B v1 → EvoQwen2.5 VL Retriever 3B v1
Re-embed your whole corpus›
Vectors from different models can't be mixed; dimensions change from 3,584 to 2,048.