Leaderboard
Multi-dimensional rankings based on model speed tests and provider health checks. Compare providers, endpoints, and reliability at a glance.
Ranked by median time to first token (resistant to outliers). Lower is better for responsiveness.
| Rank | Provider | Model | First Token Latency | Avg tokens per second | Total Tests |
|---|---|---|---|---|---|
1NEW | 0.24 s Best: 0.21Worst: 0.31 | 67.44t/s | 10 | ||
2NEW |
 NVIDIA NIMintegrate.api.nvidia.com |
| qwen/qwen3.5-122b-a10b |
0.26 s Best: 0.25Worst: 0.37 |
87.95t/s |
5 |
3NEW |  SiliconFlowapi.siliconflow.cn | THUDM/GLM-4-9B-0414 | 0.31 s Best: 0.29Worst: 0.84 | 58.52t/s | 5 |
43 | A AI Toolsplatform.aitools.cfd | zhipu/glm-4v-flash | 0.35 s Best: 0.34Worst: 0.43 | 52.78t/s | 5 |
5NEW | NVIDIA NIMintegrate.api.nvidia.com | minimaxai/minimax-m2.5 | 0.49 s Best: 0.23Worst: 5.89 | 59.50t/s | 10 |
6NEW | A AI Toolsplatform.aitools.cfd | qwen/qwen2.5-7b | 0.59 s Best: 0.35Worst: 2.00 | 95.64t/s | 5 |
77 | A AI Toolsplatform.aitools.cfd | zhipu/glm-4-flash | 0.65 s Best: 0.50Worst: 1.19 | 32.56t/s | 400 |
8NEW |  阿里云百炼 DashScopedashscope.aliyuncs.com | qwen-math-turbo | 0.65 s Best: 0.62Worst: 2.48 | 48.90t/s | 5 |
9NEW | 温 温云sxtuyxrxcgim.ap-northeast-1.clawcloudrun.com | moonshotai/kimi-k2-instruct-0905 | 0.74 s Best: 0.70Worst: 1.10 | 65.70t/s | 10 |
10NEW | Q QYES AIai.qyes.top | GLM-4-Flash-250414 | 0.74 s Best: 0.58Worst: 1.39 | 38.40t/s | 10 |
113 | A AI Toolsplatform.aitools.cfd | qwen/qwen2.5-7b | 0.75 s Best: 0.40Worst: 2.65 | 62.06t/s | 10 |
12NEW | X XShuLab Sub2APIapi.xshulab.com | gpt-5.4-mini | 0.80 s Best: 0.63Worst: 1.54 | 162.55t/s | 10 |
13NEW | X XShuLab Sub2APIapi.xshulab.com | gpt-5.1 | 0.80 s Best: 0.76Worst: 1.39 | 153.60t/s | 5 |
14NEW |  词元流动tokenflux.dev | gpt-5.4 | 0.81 s Best: 0.71Worst: 0.85 | 46.78t/s | 5 |
15NEW |  星见雅 APIapi.xinjianya.top | 英伟达/openai/gpt-oss-120b | 0.92 s Best: 0.81Worst: 1.12 | 145.35t/s | 5 |
1611 | SiliconFlowapi.siliconflow.cn | deepseek-ai/DeepSeek-V3 | 1.05 s Best: 0.86Worst: 1.38 | 18.38t/s | 5 |
17NEW |  TokenX24tokenx24.com | gpt-5.4 | 1.05 s Best: 0.88Worst: 2.90 | 46.64t/s | 5 |
18NEW | NVIDIA NIMintegrate.api.nvidia.com | qwen/qwen3-coder-480b-a35b-instruct | 1.10 s Best: 0.65Worst: 6.73 | 61.02t/s | 10 |
19NEW | 阿里云百炼 DashScopedashscope.aliyuncs.com | deepseek-v3 | 1.11 s Best: 0.95Worst: 3.30 | 40.50t/s | 5 |
20NEW |  OpenRouteropenrouter.ai | openai/gpt-oss-120b:free | 1.22 s Best: 1.03Worst: 3.06 | 33.62t/s | 5 |
21NEW |  Sub2APIapi.243706.xyz | gpt-5.4 | 1.23 s Best: 0.88Worst: 2.98 | 49.20t/s | 25 |
22NEW | 星见雅 APIapi.xinjianya.top | openai/gpt-oss-120b | 1.26 s Best: 1.05Worst: 1.37 | 160.02t/s | 5 |
23NEW |  OpenCodeopencode.ai | trinity-large-preview-free | 1.32 s Best: 1.04Worst: 1.51 | 5.44t/s | 5 |
24NEW | OpenRouteropenrouter.ai | arcee-ai/trinity-large-preview:free | 1.33 s Best: 0.96Worst: 1.74 | 5.99t/s | 10 |
25NEW | 星见雅 APIapi.xinjianya.top | grok-4.20-0309 | 1.34 s Best: 1.03Worst: 1.59 | 114.13t/s | 5 |
26NEW | 星见雅 APIapi.xinjianya.top | deepseek-chat | 1.40 s Best: 1.31Worst: 1.69 | 42.55t/s | 5 |
27NEW | 阿里云百炼 DashScopecoding.dashscope.aliyuncs.com | kimi-k2.5 | 1.40 s Best: 1.07Worst: 2.60 | 35.51t/s | 5 |
28NEW | X XShuLab Sub2APIapi.xshulab.com | gpt-5.3-codex-spark | 1.58 s Best: 0.91Worst: 2.01 | 20.79t/s | 5 |
29NEW | T Thorbasedashboard.thorbase.com | deepseek/deepseek-v3.2 | 1.67 s Best: 1.39Worst: 3.13 | 66.04t/s | 5 |
307 |  MiniMaxapi.minimaxi.com | MiniMax-M2.7-highspeed | 1.80 s Best: 1.37Worst: 3.45 | 59.37t/s | 5 |
31NEW | M MapleLeaf APIai.071129.xyz | anthropic/claude-6-opus | 1.83 s Best: 1.77Worst: 2.73 | 50.40t/s | 5 |
32NEW | A AI Toolsplatform.aitools.cfd | openai/gpt-oss-20b | 1.93 s Best: 1.51Worst: 2.66 | 42.54t/s | 5 |
33NEW |  七牛云api.qnaigc.com | openai/gpt-5.4 | 1.96 s Best: 1.91Worst: 2.14 | 53.15t/s | 5 |
34NEW | 丰 丰思理 AIai.fengsili.online | kilo/trinity-large-thinking | 1.97 s Best: 1.66Worst: 2.17 | 108.65t/s | 5 |
35NEW | X XShuLab Sub2APIapi.xshulab.com | gpt-5.4 | 2.00 s Best: 1.00Worst: 5.05 | 21.11t/s | 15 |
36NEW | 云智APIyunzhiapi.cn | DeepSeek-V3.2-EXP | 2.03 s Best: 1.64Worst: 2.50 | 34.57t/s | 15 |
37NEW | G GuaiHubguaihub.com | gpt-5.4-fast | 2.11 s Best: 1.80Worst: 4.54 | 67.71t/s | 5 |
38NEW | M Mars HKmars-hk.duckdns.org:38317 | gpt-5.3-codex-spark | 2.16 s Best: 1.17Worst: 2.99 | 184.15t/s | 14 |
39NEW | MiniMaxapi.minimaxi.com | MiniMax-M2.7 | 2.32 s Best: 1.66Worst: 4.43 | 38.92t/s | 5 |
40NEW | M Mars HKmars-hk.duckdns.org:38317 | gpt-5.4 | 2.67 s Best: 1.78Worst: 3.83 | 68.05t/s | 15 |
41NEW | w wzjself中转站wzjself.org | gpt-5.4 | 3.09 s Best: 3.00Worst: 3.95 | 21.06t/s | 5 |
42NEW | M Mars HKmars-hk.duckdns.org:38317 | gpt-5.4(xhigh) | 4.76 s Best: 4.04Worst: 7.25 | 75.46t/s | 5 |
43NEW | w wzjself中转站wzjself.org | gpt-5.4-mini | 4.98 s Best: 2.07Worst: 7.31 | 86.59t/s | 10 |
44NEW |  9Routerrb6k9jv.9router.com | NL | 5.48 s Best: 3.41Worst: 16.67 | 66.05t/s | 8 |
45NEW | 星见雅 APIapi.xinjianya.top | gpt-5.3-codex | 5.80 s Best: 4.26Worst: 14.46 | 22.24t/s | 10 |
46NEW | E EasyMoreai.easymoreapi.com | gpt-5-chat | 6.14 s Best: 4.80Worst: 9.14 | 47.09t/s | 5 |
47NEW |  DeepSeekapi.deepseek.com | deepseek-reasoner | 6.52 s Best: 5.61Worst: 9.79 | 30.91t/s | 5 |
48NEW | OpenCodeopencode.ai | gpt-5-nano | 6.60 s Best: 4.39Worst: 9.30 | 112.17t/s | 5 |
49NEW | A AI Toolsplatform.aitools.cfd | zhipu/glm-4.6v-flash | 6.65 s Best: 3.78Worst: 24.70 | 72.10t/s | 5 |
50NEW | 初 初叶🍂Furry APIai.chuyel.top | grok-4.20-beta | 6.74 s Best: 5.17Worst: 26.91 | 85.78t/s | 5 |
