Leaderboard
Multi-dimensional rankings based on model speed tests and provider health checks. Compare providers, endpoints, and reliability at a glance.
Average tokens generated per second. Higher is better for fast responses.
| Rank | Provider | Model | Throughput | Avg first token latency | Total Tests |
|---|---|---|---|---|---|
| 1 | claude37_sonnet | 2766.82 t/s Best: 4454.36Worst: 1066.47 | 0.59s | 5 | |
| 2 | deepseek/deepseek-v3-0324 | 456.45 t/s Best: 2614.87Worst: 23.04 | 7.40s | 70 |
| 3 | A AI Toolsplatform.aitools.cfd | deepseek/deepseek-v3 | 372.18 t/s Best: 2595.09Worst: 26.41 | 6.81s | 25 |
| 4 | New APIapi.lianwusuoai.top | Qwen/Qwen2-1.5B-Instruct | 213.84 t/s Best: 251.60Worst: 195.56 | 0.67s | 5 |
| 5 | ChatAnywhereapi.chatanywhere.tech | gpt-4.1-nano | 208.55 t/s Best: 273.12Worst: 123.60 | 0.65s | 5 |
| 6 | New APIapi.lianwusuoai.top | Pro/Qwen/Qwen2-1.5B-Instruct | 204.04 t/s Best: 207.89Worst: 196.17 | 0.60s | 5 |
| 7 | New APIapi.lianwusuoai.top | 免费Qwen2-1.5B | 203.36 t/s Best: 211.26Worst: 197.76 | 0.68s | 5 |
| 8 | ai.api.xn--fiqs8sai.api.xn--fiqs8s | gemini-2.0-flash | 194.65 t/s Best: 230.29Worst: 170.03 | 0.95s | 15 |
| 9 | Z Zhongzhuan Chatapi.zhongzhuan.chat | gemini-2.0-flash | 190.98 t/s Best: 211.82Worst: 163.01 | 0.81s | 5 |
| 10 | New APIapi.lianwusuoai.top | 免费Grok3-mini | 180.93 t/s Best: 198.38Worst: 147.42 | 3.99s | 5 |
| 11 | ZetaTechs APIapi.zetatechs.com | gemini-2.0-flash-lite | 157.78 t/s Best: 179.54Worst: 134.50 | 1.20s | 5 |
| 12 | Mistral AImistral.ai | codestral-latest | 151.00 t/s Best: 172.50Worst: 113.42 | 0.38s | 5 |
| 13 | SiliconFlowapi.siliconflow.cn | deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B | 142.95 t/s Best: 153.18Worst: 126.51 | 4.47s | 5 |
| 14 | New APIapi.lianwusuoai.top | deepseek-ai/deepseek-vl2 | 125.77 t/s Best: 146.34Worst: 65.75 | 0.75s | 5 |
| 15 | 智谱AI开放平台open.bigmodel.cn | glm-z1-flash | 116.46 t/s Best: 118.18Worst: 114.94 | 0.43s | 5 |
| 16 | YUNWU APIyunwu.ai | gpt-4o-mini | 114.75 t/s Best: 162.91Worst: 68.49 | 8.66s | 5 |
| 17 | N New APItranslate-api.665.pp.ua | translate-model-fast | 98.89 t/s Best: 104.36Worst: 83.70 | 0.97s | 5 |
| 18 | 小波 APIiai.iisbo.com | [bo]gemini-2.5-pro-exp-03-25 | 97.92 t/s Best: 110.90Worst: 86.08 | 14.48s | 5 |
| 19 | New APIapi.lianwusuoai.top | Pro/Qwen/Qwen2-7B-Instruct | 96.25 t/s Best: 102.85Worst: 86.93 | 0.63s | 5 |
| 20 | New APIapi.lianwusuoai.top | deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B | 94.25 t/s Best: 172.81Worst: 41.52 | 5.51s | 5 |
| 21 | New APIapi.lianwusuoai.top | Qwen/Qwen2-7B-Instruct | 93.64 t/s Best: 101.48Worst: 82.69 | 0.68s | 5 |
| 22 | SiliconFlowapi.siliconflow.cn | Qwen/Qwen2-7B-Instruct | 93.55 t/s Best: 100.52Worst: 83.83 | 0.64s | 5 |
| 23 | New APIapi.lianwusuoai.top | Pro/Qwen/Qwen2-VL-7B-Instruct | 93.23 t/s Best: 98.94Worst: 85.28 | 0.71s | 5 |
| 24 | New APIapi.lianwusuoai.top | 免费Qwen2-7B | 93.01 t/s Best: 98.84Worst: 85.80 | 0.85s | 10 |
| 25 | New APIapi.lianwusuoai.top | 免费DS-VL2 | 86.80 t/s Best: 140.48Worst: 31.60 | 1.86s | 5 |
| 26 | New APIapi.lianwusuoai.top | 免费Grok3 | 85.51 t/s Best: 116.83Worst: 66.02 | 1.39s | 5 |
| 27 | New APIapi.lianwusuoai.top | Pro/Qwen/Qwen2.5-VL-7B-Instruct | 85.51 t/s Best: 96.72Worst: 74.62 | 0.76s | 5 |
| 28 | SiliconFlowapi.siliconflow.cn | THUDM/GLM-Z1-9B-0414 | 78.80 t/s Best: 80.11Worst: 77.60 | 8.89s | 5 |
| 29 | 1 117.72.69.228:32576117.72.69.228:32576 | QwQ-32B | 78.79 t/s Best: 81.51Worst: 77.55 | 8.16s | 10 |
| 30 | SiliconFlowapi.siliconflow.cn | Qwen/Qwen3-14B | 78.64 t/s Best: 83.29Worst: 73.98 | 9.81s | 5 |
| 31 | New APIapi.lianwusuoai.top | 免费Qwen2.5-14B | 77.95 t/s Best: 80.57Worst: 74.32 | 0.69s | 5 |
| 32 | New APIapi.lianwusuoai.top | Qwen/Qwen2.5-14B-Instruct | 77.92 t/s Best: 81.71Worst: 68.44 | 0.69s | 5 |
| 33 | New APIapi.lianwusuoai.top | 免费Qwen2-VL-7B | 77.82 t/s Best: 96.06Worst: 52.43 | 0.95s | 10 |
| 34 | New APIapi.lianwusuoai.top | internlm/internlm2_5-7b-chat | 73.88 t/s Best: 80.00Worst: 63.15 | 0.61s | 5 |
| 35 | New APIapi.lianwusuoai.top | 免费GLM-4-9B-128K | 73.70 t/s Best: 76.98Worst: 71.34 | 0.77s | 5 |
| 36 | New APIapi.lianwusuoai.top | Qwen/QwQ-32B-Preview | 72.57 t/s Best: 74.00Worst: 71.41 | 0.81s | 5 |
| 37 | New APIapi.lianwusuoai.top | THUDM/glm-4-9b-chat | 71.17 t/s Best: 77.40Worst: 63.54 | 0.65s | 5 |
| 38 | New APIapi.lianwusuoai.top | Qwen/QwQ-32B | 69.22 t/s Best: 86.98Worst: 37.41 | 14.67s | 5 |
| 39 | New APIapi.lianwusuoai.top | Pro/THUDM/glm-4-9b-chat | 68.04 t/s Best: 80.64Worst: 60.18 | 0.87s | 5 |
| 40 | SiliconFlowapi.siliconflow.cn | Qwen/Qwen2.5-7B-Instruct | 67.51 t/s Best: 68.74Worst: 65.56 | 0.87s | 5 |
| 41 | SiliconFlowapi.siliconflow.cn | deepseek-ai/DeepSeek-R1-Distill-Qwen-7B | 66.31 t/s Best: 76.96Worst: 48.82 | 12.96s | 5 |
| 42 | New APIapi.lianwusuoai.top | 免费Qwen2.5-VL-7B | 63.63 t/s Best: 76.63Worst: 45.85 | 1.03s | 10 |
| 43 | New APIapi.lianwusuoai.top | internlm/internlm2_5-20b-chat | 61.81 t/s Best: 71.71Worst: 38.91 | 0.77s | 5 |
| 44 | A AI Toolsplatform.aitools.cfd | deepseek/deepseek-r1 | 61.18 t/s Best: 437.13Worst: 0.00 | 11.03s | 60 |
| 45 | New APIapi.lianwusuoai.top | Qwen/Qwen2.5-32B-Instruct | 60.92 t/s Best: 72.45Worst: 53.04 | 0.86s | 5 |
| 46 | a api.centml.comapi.centml.com | deepseek-ai/DeepSeek-V3-0324 | 59.16 t/s Best: 71.59Worst: 48.49 | 0.67s | 5 |
| 47 | ChatAnywhereapi.chatanywhere.tech | gpt-4o-mini | 57.47 t/s Best: 112.18Worst: 29.83 | 2.08s | 5 |
| 48 | A AI Toolsplatform.aitools.cfd | deepseek/deepseek-r1-32b | 55.71 t/s Best: 57.01Worst: 53.83 | 3.68s | 5 |
| 49 | SiliconFlowapi.siliconflow.cn | Qwen/Qwen2.5-32B-Instruct | 55.30 t/s Best: 64.22Worst: 52.39 | 0.81s | 5 |
| 50 | A AI Toolsplatform.aitools.cfd | google/gemma-3-27b | 54.54 t/s Best: 93.12Worst: 4.62 | 2.63s | 20 |