Kimi K3.1 Pro Gains Day-0 Support Across Inference Platforms
Moonshot's Kimi K3.1 Pro, its strongest open-weight model, launched with Day-0 support from You.com (1M context), Tinker (256K context), and SGLang (33.7K tok/s per GPU via sparse attention), with SemiAnalysis benchmarks showing 45 million tokens per dollar on MI355X.
2026-11-05 ~ 2026-11-07 · 4 related posts
- Kimi K3.1 Pro live on You.com Inference Router with 1M context, 65.7% Terminal-Bench — PolarBearby · 2026-11-05
- Kimi K3.1 Pro Day 0: SemiAnalysis TCO chart shows MI355X beating B200/B300 on tokens per dollar — AnushElangovan · 2026-11-07
- Kimi K3.1 Pro lands on Tinker with 256K context, its largest open-weights model — _kevinlu · 2026-11-07
- SGLang ships Day-0 support for Kimi K3.1 Pro with sparse attention, up to 33.7K tok/s per GPU — ying11231 · 2026-11-07