FLUX.1 [dev]
BFL:FLUX-1-DEVImage DiTOperationalLoad:45%SLA:99.98%Zero Queue
12B Rectified Flow Transformer
Base Rate
24 Credits / run
24h +18.42%
Flagship open-weight 12B flow transformer by Black Forest Labs, leading text-to-image realism and typography adherence.
24h Total Calls
3.84Mcalls
+18.42% vs yesterday
Total Generations
3.12Mimgs
+16.8% · 4K Ratio 34%
Avg Latency (P50)
2.4s
20-step 2.1s · 4K Hi-res 3.8s
Availability SLA
99.98%
Failover <85ms
Model Telemetry & Volume Chart
BFL:FLUX-1-DEVInvocations
168,092次+8.38%
Live Telemetry
Operational Performance
Production-grade Latency & Concurrency
P50 TTFT210 ms
P99 Latency3850 ms
Throughput TPS
Peak Concurrency1,420 req
Packet Loss Rate0.001%
Cost & Billing Matrix
Credits and Token Breakdown
Base Task Cost24 Credits
Input Cost / 1k Tokens¥0.012
Output Cost / 1k Tokens¥0.036
4K Upscale Multiplier1.6x
VIP Discount至高 25% 优惠
Technical Specifications
Parameters, Context & Resolutions
Architecture12B Rectified Flow Transformer
Parameters12 Billion
Context Window4,096 tokens
Max Resolution2048 x 2048 (4K Upscaled)
Max DurationN/A
Global Edge Gateways
Live Region Latency & Load Balancing
East China (Shanghai)
18ms
South China (Guangzhou)
22ms
Hong Kong Gateway
32ms
Asia Pacific (Singapore)
55ms
US West (Silicon Valley)
142ms
EU Central (Frankfurt)
165ms
Comparable Models in Same Domain
Models with similar creative modalities
Seedream 5.0 (即梦)
BYTEDANCE:SEEDREAM-5.0
ByteDance flagship creative visual foundation model with 4K cinema-grade rendering and oriental aesthetic mastery.
Rate28 Credits / run
GPT-Image 2.5 Flash
OPENAI:GPT-IMG-2.5F
Sub-second image generation engine by OpenAI, tailored for conversational ideation and strict prompt instruction following.
Rate18 Credits / run
通义万相 2.1 (Wan 2.1)
ALIBABA:WAN-2.1-T2I
Alibaba flagship bilingual visual transformer excelling in Chinese typography, commercial poster synthesis, and realistic lighting.
Rate20 Credits / run