← TPUs, AI chips & CPU endpoints
Compare 3 tpus, ai chips & cpu endpoints APIs
| HF Endpoints Google TPU Hugging Face Inference Endpoints | HF Endpoints CPU (4 vCPU) Hugging Face Inference Endpoints | HF Endpoints AWS Inferentia2 Hugging Face Inference Endpoints | |
|---|---|---|---|
| $ / accel-hr | $1.20 Verified Oct 6, 2026 | $0.134 Verified Oct 6, 2026 | $0.75 Verified Oct 6, 2026 |
| Quality | – | – | – |
| Speed | – | – | – |
| Accelerator type | tpu | cpu | inferentia |
| Memory | — | 8 GB | — |
| Frameworks | jax, pytorch | onnx, pytorch | pytorch |
| Supports training | No | No | No |
| Per-minute billing | Yes | Yes | Yes |