Sarvam's new 105B checkpoint jumps from 42.6 to 73.9 on benchmarks, claims Flash-Lite Pareto frontier
itsOmSarraf_ · x · 2026-11-22
Indian AI lab Sarvam AI released Sarvam-105B-1711, a new checkpoint of its 105B model in the Flash-Lite category, claiming a new Pareto frontier averaged across instruction following, chat, tool calling, and Indian languages.
The accompanying price-performance chart shows the benchmark average leaped from 42.6 (March's Sarvam-105B) to 73.9, at a blended price around $0.15 per million tokens — sitting on the Pareto frontier against Qwen3.8 27B (74.6), Gemini 3.5 Flash-Lite (74.0), Gemma 4 31B (73.0), GPT-6 Luna (66.3), Nemotron 3 Super (62.2), and Mistral Small 4 (49.7).
More from Models
- Dev says Google models are making real progress on lossless compression as intelligence proxy — gnukeith · 2026-11-22
- Astra-6.2 'lazy' backlash blamed on user education: default manual mode vs goal mode — alexisgallagher · 2026-11-22
- Musk: Tesla has had a near pixel-perfect world model since last year — giffmana · 2026-11-22
- Ant open-sources LingSante-3.1-flash, a 560B healthcare model with FP8/MXFP4 weights — BerryBlush_Rogers691 · 2026-11-22
- Gemini 4 Neon takes #1 on LiveBench 2.0 with 87.4, beating Claude and GPT by 2.5 points — demishassabis · 2026-11-22
- User Reports Noticeable Quality Drop in Claude Opus 5.5 Recently — jiayuan_jy · 2026-11-22