Qwen 2.5 3B
On iPhone 16 Pro Max, 8 GB with iOS 27.2, Nov 3, 2026. Run by Toskar.
Runs: Fully
Toskar Score119
How fast it ran
- Writes
- 21.7 tokens a second
- Reads
- 187 tokens a second
- First token, with a 4,269-token prompt
- 22.9 s
- Start to finish
- 2.6 tokens a second
- On the GPU
- 37 of 37 layers
- Quantization
- Q4_K_M
- Context window
- 8,192 tokens
Method version 1. One result, on one computer. It isn't a ranking, and scores from other computers or versions aren't the same test.
Compare with other computersHow this is measuredThink this is wrong? Tell usThis runCan my computer run Qwen 2.5 3B?