Sarvam open-sources Prism, a 44ms decision model built on Sarvam 105B
itsOmSarraf_ · x · 2026-11-03
Sarvam AI released and open-sourced Sarvam Prism, a System 1 decision model built on Sarvam 105B for fast, accurate decisions in enterprise workflows.
- Scores 71.8 with a median response time of 44ms — the highest average score among decision models responding under 100ms
- Uses block-bidirectional attention, isolates options via masks and shared positions, and a split attention kernel for fast inference
- The team claims it pushes the pareto frontier of accuracy vs. speed vs. calibration
Related event: India's Sarvam Open-Sources Prism Decision Model with 44ms Response(2 posts)→
More from Models
- 'Astra can rewrite Qwen weights' hits 2863 points on HN, then a rebuttal calls it misleading — yoshiyama_akira · 2026-11-03
- LLMs can only read explicit vibes, not implicit ones, VC Stewart Alsop III argues — StewartalsopIII · 2026-11-03
- DeepSeek-V4.1-Flash GGUF quantized build trends on Hugging Face — ISTA-DASLab · 2026-11-03
- Qwen3.5-4B fine-tune for 3D/GLB code generation trends on Hugging Face — ccore · 2026-11-03
- llama.cpp Merges Support for NVIDIA's Nemotron 3.5 Transcribe: Offline ASR in 40 Languages — jacek2023 · 2026-11-03
- Blogger slams Opus 5.5 medium as lazy, says astra figures things out on its own — justalexoki · 2026-11-03