All models
DeepSeek: DeepSeek V4.1 Flash 1017
deepseek/deepseek-v4.1-flash-1017 · 1,048,576 context · 30 providers listed
Lowest listed
$0.022
OpenInference
Mid listed
$0.381
middle of every listing
Gap
17.2×
94% off at the lowest
Hosts, lowest first
The five lowest of 30 hosts carrying this model. Per million tokens, compared as three parts input to one part output.
| Host | Precision | Weighted | Input | Output | Context |
|---|---|---|---|---|---|
| OpenInferencelowest | fp4 | $0.022 | $0.020 | $0.029 | 1,048,576 |
| Sail Research | fp4 | $0.113 | $0.101 | $0.150 | 1,048,576 |
| DeepInfra | fp8 | $0.118 | $0.078 | $0.235 | 1,048,576 |
| Reka | unknown | $0.155 | $0.040 | $0.500 | 1,048,576 |
| Relace | unknown | $0.178 | $0.070 | $0.500 | 1,048,576 |
The same weights, at very different prices
A listing says what a host charges, not whether it carries the real weights or bills what it lists. Our proof runs put the same questions to every host, compare the answers, then count the tokens billed. Hosts that hold up are proven, and only proven hosts see your requests.