FreeLLMAll models › Compare
DeepSeek-V4-Pro (High, 20260813) vs MiMo-V2.5-Pro
MiMo-V2.5-Pro is ahead on Arena (1465 vs 1451): in a blind vote people prefer it about 52% of the time.
DeepSeek-V4-Pro (High, 20260813)1451Arena chat · #49 of 416
MiMo-V2.5-Pro1465Arena chat · #37 of 416
Arena scores by category
| Category | DeepSeek-V4-Pro (High, 20260813) | MiMo-V2.5-Pro | Ahead |
|---|---|---|---|
| Chat | 1451 | 1465 | MiMo-V2.5-Pro |
| Coding | 1582 | 1481 | DeepSeek-V4-Pro (High, 20260813) |
Scores from blind human votes on Arena. A gap under 10 points is too small to call.
Size and where to use them free
| DeepSeek-V4-Pro (High, 20260813) | MiMo-V2.5-Pro | |
|---|---|---|
| Made by | deepseek | xiaomi |
| Size | 1598.8 B (75.6 B active) | 1023.2 B (43.8 B active) |
| 4-bit download | 970 GB | 621 GB |
| Use it free | download and run at home · MIT license | download and run at home · MIT license |
Speed at home
| Hardware | DeepSeek-V4-Pro (High, 20260813) | MiMo-V2.5-Pro |
|---|---|---|
| RTX 3090 | does not fit | does not fit |
| RTX 4090 | does not fit | does not fit |
| RTX 5090 | does not fit | does not fit |
| RX 7900 XTX | does not fit | does not fit |
| Mac M4 Max 128 GB | does not fit | does not fit |
| Ryzen AI Max+ 395 128 GB | does not fit | does not fit |
Estimates at 4-bit with an 8K context, 64 GB of system RAM for offload, calibrated on 44 real llama.cpp tests.
Try them yourself
Ask both the same question side by side in Try & compare, or vote blind between free models in Free Arena.
More comparisons
- DeepSeek-V4-Pro (High, 20260813) vs MiMo-V2.6-Pro
- DeepSeek-V4-Pro (High, 20260813) vs Kimi K3 (Max)
- DeepSeek-V4-Pro (High, 20260813) vs GLM-5.3 (Max)
- DeepSeek-V4-Pro (High, 20260813) vs GLM-5.3-Flash
- DeepSeek-V4-Pro (High, 20260813) vs GLM-5.2 (Max)
- DeepSeek-V4-Pro (High, 20260813) vs deepseek-v4.1-flash-max
- MiMo-V2.5-Pro vs MiMo-V2.6-Pro
- MiMo-V2.5-Pro vs Kimi K3 (Max)
- MiMo-V2.5-Pro vs GLM-5.3 (Max)
- MiMo-V2.5-Pro vs GLM-5.3-Flash