Live Model Latency & Throughput Benchmarks
Real-time 24-hour Time to First Token (TTFT), generation speed (tokens/sec), and reliability across active models.
Active models
43
Fastest TTFT (24h)
2,171 ms
Peak speed (24h)
526 tok/s
| Model | Official provider | Avg TTFT | Output speed | 24h success | 24h requests |
|---|---|---|---|---|---|
| Gemini 2.5 Flash Image gemini-2.5-flash-image | 2,171 ms | 161 tok/s | 100% | 5 | |
| DeepSeek V4.1 Flash deepseek-v4.1-flash | DeepSeek | 2,203 ms | 320 tok/s | 100% | 24,911 |
| GLM 5.3 Flash glm-5.3-flash | Z.ai | 2,243 ms | 37 tok/s | 100% | 213 |
| DeepSeek V4 Pro deepseek-v4-pro | DeepSeek | 2,285 ms | 85 tok/s | 100% | 81 |
| Claude Opus 5.5 claude-opus-5-5 | Anthropic | 2,687 ms | 208 tok/s | 100% | 282 |
| DeepSeek V4 Flash 0731 deepseek-v4-flash-0731 | DeepSeek | 3,651 ms | 64 tok/s | 100% | 81 |
| GPT 6 Luna gpt-6-luna | OpenAI | 12,364 ms | 526 tok/s | 97% | 36 |
| MiMo V2.5 mimo-v2.5 | Xiaomi | 20,653 ms | 46 tok/s | 100% | 28 |
| Claude Fable 5 claude-fable-5 | Anthropic | Awaiting 24h sample | — | — | — |
| Claude Fable 5.1 claude-fable-5-1 | Anthropic | Awaiting 24h sample | — | — | — |
| Claude Opus 4.6 claude-opus-4.6 | Anthropic | Awaiting 24h sample | 20 tok/s | 100% | 16 |
| Claude Opus 4.7 claude-opus-4-7 | Anthropic | Awaiting 24h sample | — | — | — |
| Claude Opus 4.8 claude-opus-4-8 | Anthropic | Awaiting 24h sample | — | — | — |
| Claude Sonnet 5.5 claude-sonnet-5-5 | Anthropic | Awaiting 24h sample | — | — | — |
| DeepSeek V4 Flash deepseek-v4-flash | DeepSeek | Awaiting 24h sample | — | — | — |
| DeepSeek V4 Pro 0813 deepseek-v4-pro-0813 | DeepSeek | Awaiting 24h sample | — | — | — |
| DeepSeek V4.1 Flash Fast deepseek-v4.1-flash-fast | DeepSeek | Awaiting 24h sample | — | — | — |
| Gemini 3.1 Pro gemini-3.1-pro | Awaiting 24h sample | — | — | — | |
| Gemini 3.8 Flash gemini-3.8-flash | Awaiting 24h sample | — | — | — | |
| Gemini 4 Argon gemini-4-argon | Awaiting 24h sample | — | — | — | |
| GLM 5.2 glm-5.2 | Z.ai | Awaiting 24h sample | — | — | — |
| GLM 5.3 glm-5.3 | Z.ai | Awaiting 24h sample | — | — | — |
| GLM 5.3 FlashX glm-5.3-flashx | Z.ai | Awaiting 24h sample | — | — | — |
| GPT 5.6 Luna gpt-5.6-luna | OpenAI | Awaiting 24h sample | — | — | — |
| GPT 5.6 Sol gpt-5.6-sol | OpenAI | Awaiting 24h sample | — | — | — |
| GPT 5.6 Terra gpt-5.6-terra | OpenAI | Awaiting 24h sample | — | — | — |
| GPT 6 Astra gpt-6-astra | OpenAI | Awaiting 24h sample | — | — | — |
| GPT 6.1 Sol gpt-6.1-sol | OpenAI | Awaiting 24h sample | — | — | — |
| GPT Image 2 gpt-image-2 | OpenAI | Awaiting 24h sample | — | — | — |
| Grok 4.6 grok-4.6 | xAI | Awaiting 24h sample | — | — | — |
| Grok 4.7 grok-4.7 | xAI | Awaiting 24h sample | — | 100% | 1 |
| Kimi K3 kimi-k3 | Moonshot AI | Awaiting 24h sample | — | — | — |
| Kimi K3 (256k) kimi-k3-256k | Moonshot AI | Awaiting 24h sample | — | — | — |
| Kimi K3.1 Pro kimi-k3.1-pro | Moonshot AI | Awaiting 24h sample | — | — | — |
| MiMo V2.5 Pro mimo-v2.5-pro | Xiaomi | Awaiting 24h sample | — | — | — |
| MiniMax M3 minimax-m3 | MiniMax | Awaiting 24h sample | — | — | — |
| Qwen3.7 Max qwen3.7-max | Alibaba Qwen | Awaiting 24h sample | — | — | — |
| Qwen3.7 Plus qwen3.7-plus | Alibaba Qwen | Awaiting 24h sample | — | — | — |
| Qwen3.8 27B qwen3.8-27b | Alibaba Qwen | Awaiting 24h sample | — | — | — |
| Qwen3.8 Flash qwen3.8-flash | Alibaba Qwen | Awaiting 24h sample | — | — | — |
| Qwen3.8 Max qwen3.8-max | Alibaba Qwen | Awaiting 24h sample | — | — | — |
| Qwen3.8 Max 0902 qwen3.8-max-0902 | Alibaba Qwen | Awaiting 24h sample | — | — | — |
| Qwen3.8 Omni Flash qwen3.8-omni-flash | Alibaba Qwen | Awaiting 24h sample | — | — | — |