Select runtimes to compare side by side. Click chips below to toggle selection.
| Metric | Wasmtime | Python (CPython) | CUDA Runtime | Node.js | llama.cpp | Docker | KoboldCpp |
|---|---|---|---|---|---|---|---|
| Score | A88 | B71 | B-65 | B-65 | C+63 | C-54 | F38 |
| Type | WebAssembly | Language | Language | Container | |||
| Execution | aot | interpreted | aot | jit | aot | hybrid | hybrid |
| Interface | cli | cli | sdk | cli | cli | cli | gui |
| Cold Start | 1ms | 50ms | 100ms | 50ms | 100ms | 500ms | 1500ms |
| Memory | 5MB | 15MB | 500MB | 40MB | 50MB | 50MB | 400MB |
| Startup | <1ms | 10ms | 50ms | 20ms | 10ms | 200ms | 300ms |
| Isolation | process | process | hardware | process | process | container | process |
| Maturity | production | production | production | production | production | production | stable |
| Languages | Rust, C, C++, Go, Python, AssemblyScript | Python | C, C++, Python | JavaScript, TypeScript | C, C++ | Any | C++, Python |
| License | Apache-2.0 | Other | Proprietary | MIT | MIT | Apache-2.0 | AGPL-3.0 |
| Links |