FreeLLMAll models › Compare
DeepSeek-V4.1-Flash-20260910 (Max) vs Kimi K2.5 (Thinking)
DeepSeek-V4.1-Flash-20260910 (Max) is ahead on Arena (1476 vs 1450): in a blind vote people prefer it about 54% of the time.
DeepSeek-V4.1-Flash-20260910 (Max)1476Arena chat · #35 of 419
Kimi K2.5 (Thinking)1450Arena chat · #80 of 419
Arena scores by category
| Category | DeepSeek-V4.1-Flash-20260910 (Max) | Kimi K2.5 (Thinking) | Ahead |
|---|---|---|---|
| Chat | 1476 | 1450 | DeepSeek-V4.1-Flash-20260910 (Max) |
| Coding | 1619 | 1436 | DeepSeek-V4.1-Flash-20260910 (Max) |
| Vision | 1261 | 1252 | about even |
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-V4.1-Flash-20260910 (Max) | Kimi K2.5 (Thinking) | |
|---|---|---|
| Made by | deepseek | moonshot |
| Size | not public | 1026.9 B (32.2 B active) |
| 4-bit download | — | 624 GB |
| Use it free | MIT license | download and run at home · Modified MIT license |
Speed at home
| Hardware | DeepSeek-V4.1-Flash-20260910 (Max) | Kimi K2.5 (Thinking) |
|---|---|---|
| RTX 3090 | online only | does not fit |
| RTX 4090 | online only | does not fit |
| RTX 5090 | online only | does not fit |
| RX 7900 XTX | online only | does not fit |
| Mac M4 Max 128 GB | online only | does not fit |
| Ryzen AI Max+ 395 128 GB | online only | 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.
Share this page:XLinkedInRedditWhatsApp
Spotted a wrong number or a broken link? Report a mistake