AI Arena
Two models, same prompt, blind vote. You decide which AI is better.
Enter a prompt and hit Battle to start
Models are chosen randomly and hidden
Leaderboard
| # | Model | ELO | Wins | Losses | Ties | Win Rate |
|---|---|---|---|---|---|---|
| 1 | DeepSeek R1 7B (retired) | 1248 | 37 | 39 | 41 | 32% |
| 2 | Kimi K2.5 | 1216 | 1 | 0 | 0 | 100% |
| 3 | Deepseek Chat V3 0324 | 1215 | 1 | 0 | 1 | 50% |
| 4 | Llama 3.3 70B Instruct | 1200 | 0 | 0 | 1 | 0% |
| 5 | Gpt 4O Mini | 1200 | 1 | 1 | 0 | 50% |
| 6 | Gemini 2.5 Pro | 1184 | 0 | 1 | 0 | 0% |
| 7 | Qwen 3 8B (retired) | 1171 | 142 | 32 | 78 | 56% |
| 8 | Qwen 7B (retired) | 1158 | 35 | 123 | 77 | 15% |
| 9 | Mistral 7B (retired) | 1134 | 21 | 23 | 41 | 25% |
| 10 | Qwen 3.6 35B Thinking | 1104 | 100 | 92 | 81 | 37% |
| 11 | Qwen 3.6 35B | 1034 | 92 | 100 | 81 | 34% |
605 total votes. ELO ratings update after each vote.
Rows marked retired are results from past battles against models we no longer run. Those model ids now answer with Qwen 3.6 35B.