Kimi K3 DSpark Upgrade Boosts Long-Context & Agentic Performance, Hits 150K Downloads
BanghuaZ · x · 2026-08-01
The DSpark version of Kimi K3 has been upgraded on the SGLang framework, optimized for long-context and agentic use cases.
- Benchmark Improvements: Achieved an Accept length of 4.2 in RULER v2 1M context test and 4.66 on SWE-rebench. Improvements seen across most benchmarks (with a minor regression on GSM8K).
- Rapid Adoption: Crossed 140K downloads in just a few days post-launch and is already serving production traffic via SGLang.
Related event: Kimi K3 Upgrades Long-Context and Agent Capabilities(2 posts)→
More from Infra
- GitHub rewrote the 800k-line Copilot runtime in Rust using AI agents, cutting CPU and memory by over 90% — bibryam · 2026-10-16
- NVIDIA quietly delists the entry-level RTX 5050 from its marketplace, following the RTX 5090 — chemist_slime · 2026-10-16
- Grok claims 816B MoE model runs locally on dual M5 Ultra Mac Studios via MLX — grok · 2026-10-16
- Same 1,000 Cloudflare Durable Objects: $4,050/month cut to $22 with one trick — threepointone · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16