Global Model Real-time Telemetry
Model Screener
Real-time telemetry, invocation volumes, generation throughput, and pricing matrix across all foundation models.
24h Total Calls38,245,800+21.4%
Total Outputs26,890,200+19.8%
FLUX.1 [dev]
BFL:FLUX-1-DEV
24 Credits / run
24h Images: 3.12M imgsRender: 2.4s
Kling 2.6 Pro (可灵)
KUAISHOU:KLING-2.6-PRO
60 Credits / run
24h Videos: 1.25M vidsRender: 8.7s
Suno V3.5 Music
SUNO:V3.5-STUDIO
30 Credits / run
24h Tracks: 820k tracksTrack: 4.1s
DeepSeek-V3 671B
DEEPSEEK:V3-671B
2 Credits / 1K Tokens
24h Invocations: 12.48M reqTTFT: 110ms
Timeframe:
Showing: 12Cluster Live
| # | Model / Architecture | Type | Cluster Status | Rate | 24h Invocations | 24h Generations | Latency Breakdown | Trend | Action |
|---|---|---|---|---|---|---|---|---|---|
| 1 | DeepSeek-V3 671B DEEPSEEK:V3-671B·深度求索 (DeepSeek) | MoE LLM | 99.99%optimal Load 68% | 2 Credits / 1K Tokens | 1248.1wcalls +38.90% | 9.8BTokens Code & Reasoning 89% | 185 msTTFT 110ms · 142 tok/s | ||
| 2 |  GPT-4o Omni OPENAI:GPT-4O-OMNI·OpenAI | Omni Multimodal | 99.99%optimal Load 64% | 10 Credits / 1K Tokens | 985.0wcalls +15.20% | 742.0winferences Multimodal Alignment 96% | 220 msTTFT 140ms · 118 tok/s | ||
| 3 | FLUX.1 [dev] BFL:FLUX-1-DEV·Black Forest Labs | Image DiT | 99.98%optimal Load 45% | 24 Credits / run | 384.2wcalls +18.42% | 312.0wimgs 4K Ratio 34% | 2420 ms20-step 2.1s · 4K Hi-res 3.8s | ||
| 4 |  Seedream 5.0 (即梦) BYTEDANCE:SEEDREAM-5.0·字节跳动 (ByteDance) | Transformer Diffusion | 99.99%optimal Load 54% | 28 Credits / run | 312.1wcalls +21.30% | 268.0wimgs Native 4K Ratio 48% | 3150 msNative 4K 3.1s · Redraw 1.2s | ||
| 5 | GPT-Image 2.5 Flash OPENAI:GPT-IMG-2.5F·OpenAI | Multi-Stage Latent DiT | 99.98%optimal Load 38% | 18 Credits / run | 245.0wcalls +9.45% | 204.0wimgs Prompt Fit 99.2% | 1380 msFlash 1.3s · Prompt Fit 99% | ||
| 6 |  通义万相 2.1 (Wan 2.1) ALIBABA:WAN-2.1-T2I·阿里巴巴 (Alibaba) | DiT Multi-lingual | 99.96%optimal Load 61% | 20 Credits / run | 198.0wcalls +12.80% | 165.0wimgs Typography 98.4% | 2850 msPoster 2.8s · Typography 98% | ||
| 7 | Kling 2.6 Pro (可灵) KUAISHOU:KLING-2.6-PRO·快手 (Kuaishou) | Video DiT | 99.92%busy Load 82% | 60 Credits / run | 182.4wcalls +24.15% | 124.5wvids 1080P Ratio 82% | 8650 ms5s Video 8.6s · 60 FPS | ||
| 8 |  Midjourney V7 MJ:V7-DIRECT·Midjourney | Artistic Diffusion | 99.88%queued Load 91% | 45 Credits / run | 168.0wcalls +7.85% | 142.0wimgs Aesthetic Score 9.6 | 12500 msGrid 12.5s · Upscale 18.2s | ||
| 9 |  Hailuo 2.3 (海螺视频) MINIMAX:HAILUO-2.3·MiniMax | Video Foundation | 99.94%busy Load 75% | 55 Credits / run | 142.1wcalls +16.75% | 98.0wvids Expression 98% | 9120 ms5s Physics 9.1s · Expression 98% | ||
| 10 | Suno V3.5 Music SUNO:V3.5-STUDIO·Suno AI | Audio LLM | 99.97%optimal Load 42% | 30 Credits / run | 96.8wcalls +14.62% | 82.0wtracks Vocal Ratio 78% | 4120 msBuffer 1.2s · Full 4.1s | ||
| 11 |  Google Veo 3 GOOGLE:VEO-3-GEN·Google DeepMind | Multimodal Video | 99.85%queued Load 88% | 75 Credits / run | 89.0wcalls +8.12% | 58.0wvids Cinema Camera 88% | 12400 ms8s Cinema 12.4s · Camera Paths | ||
| 12 | MiniMax Music 3.0 MINIMAX:MUSIC-3.0·MiniMax | Audio Acoustic | 99.98%optimal Load 36% | 25 Credits / run | 62.0wcalls +11.20% | 51.0wtracks Multi-genre 94% | 3820 msBuffer 0.9s · Full 3.8s |