Try NVIDIA Llama Nemotron Ultra as an NVIDIA NIM microservice. At only 253B total parameters, it offers reasoning performance that meets or beats top open…
Try NVIDIA Llama Nemotron Ultra as an NVIDIA NIM microservice. At only 253B total parameters, it offers reasoning performance that meets or beats top open reasoning models like DeepSeek-R1 while offering considerably higher throughput due to its optimized sizing, and retaining excellent tool calling capabilities.
Source
Source:: NVIDIA