Pricing plans
Bring Triton to your chip
4-6 month service contract
For chips without Triton support
- Official Triton backend plugin
- Integrate with runtimes
- Hardware-specific code generation
- Correctness and performance validation
- Hands-on support from compiler experts
Day one model support for your chip
Yearly subscription
For chips with Triton support
- New model support
- New optimization passes
- Maintain backend plugin compatibility
- Priority support from compiler experts
- Performance regression tracking
- SLA commitments
Kernelize Platform
Free and open-source platform
Easily run inference with Triton support
- Latest Triton and LLM support
- No need to fork Triton or vLLM
- Extensible plugin architecture
- Reusable open-source building blocks
- No vendor lock-in
