Running Inkling Small on a Single Node: Native Voice Interaction Under 500ms
andimarafioti · x · 2026-07-31
A developer tested the newly released Inkling Small, highlighting that it fits on a single node (8x RTX Pro 6000 Blackwell), whereas the flagship requires 2TB of VRAM.
Plugged into HF's speech-to-speech pipeline, the model takes raw audio input directly—bypassing text transcription—and replies via faster-Qwen3TTS with an end-to-end latency under 500ms. Because the model processes native audio, it can pick up on the user's tone and emotions, showcasing strong potential for real-time multimodal interaction.
More from Infra
- NVIDIA launches RTX Spark for local AI on Windows PCs — ostrisai · 2026-10-16
- SkyRL ships Day-0 RL support for Beam with +62% LoRA decode throughput in vLLM — vllm_project · 2026-10-16
- DFlash2 delivers up to 3.03x decode-only speedup for Beam across five benchmarks — vllm_project · 2026-10-16
- vLLM's Beam serving kernels cut GB200 forward time from 8.15 to 6.47 ms at TP8 batch 128 — vllm_project · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Gemma 4 31B QAT on one 24GB card: 46 tok/s plain, ~80 with draft model on RTX 4090 — Subject_Poet_3861 · 2026-10-16