What's new
Models joining the catalogue, new coverage of ones we already knew, and changes to the reference offering.
| When | Model | Price | What | From | To | Change |
|---|---|---|---|---|---|---|
| Input (cached read) | reference changed | 0.4 | 0.3 | -25% | ||
| Output | reference changed | 4.8 | 3.6 | -25% | ||
| Input | reference changed | 1.6 | 1.2 | -25% | ||
Tencent: Hy4 previewtencent | Output | new coverage | 2.5 | 2.5 | 0% | |
Tencent: Hy4 previewtencent | Input | new coverage | 0.834 | 0.834 | 0% | |
StepFun: Step 5 PreviewStepFun | Input (cached read) | added | — | 0.05 | — | |
StepFun: Step 5 PreviewStepFun | Output | added | — | 2.7 | — | |
StepFun: Step 5 PreviewStepFun | Input | added | — | 1 | — | |
Qwen3.8 27Bnvidia | Input (cached read) | added | — | 0.04 | — | |
Qwen3.8 27Bnvidia | Output | added | — | 2.09 | — | |
Qwen3.8 27Bnvidia | Input | added | — | 0.25 | — | |
Qwen3.8 Omni Flashaihubmix | Input (cache write) | added | — | 0.1759 | — | |
Qwen3.8 Omni Flashaihubmix | Input (cached read) | added | — | 0.0141 | — | |
Qwen3.8 Omni Flashaihubmix | Output | added | — | 0.38 | — | |
Qwen3.8 Omni Flashaihubmix | Input | added | — | 0.1126 | — | |
DeepSeek V4.1 Flash Flexneuralwatt | Input (cached read) | added | — | 0.00975 | — | |
DeepSeek V4.1 Flash Flexneuralwatt | Output | added | — | 0.39 | — | |
DeepSeek V4.1 Flash Flexneuralwatt | Input | added | — | 0.0975 | — | |
Qwen3.6 35B Flexneuralwatt | Input (cached read) | added | — | 0.0189 | — | |
Qwen3.6 35B Flexneuralwatt | Output | added | — | 0.7475 | — | |
Qwen3.6 35B Flexneuralwatt | Input | added | — | 0.1885 | — | |
DeepSeek V4 Flash (Speed)neuralwatt | Input (cached read) | added | — | 0.028 | — | |
DeepSeek V4 Flash (Speed)neuralwatt | Output | added | — | 0.28 | — | |
DeepSeek V4 Flash (Speed)neuralwatt | Input | added | — | 0.14 | — | |
Qwen3.8 27B Flexneuralwatt | Input (cached read) | added | — | 0.1625 | — |