AI News · October 19, 2026

Price cuts on five LLM inputs as SGLang ships v0.5.22

69,076 stars today514 new projects+14,624 rea

Five model providers cut input prices today. Z.AI lowered GLM 5.3 Flash by 67% to $0.05 per 1M tokens. The other cuts were Qwen3 30B A3B Instruct 2507 down 10% to $0.09, Kimi K3 down 17% to $0.25, DeepSeek V4.1 Flash down 25% to $0.06, and Nemotron 3.5 Super VL down 33% to $0.1.

SGLang released v0.5.22, listing 950 PRs from 245 contributors. n8n shipped 2.43.8 with a bug fix that turns off expression spans by default. ComfyUI v0.42.1 fixed crashing issues in the MOSS image template workflow.

The index of over 10,000 open-source AI projects gained 69,076 stars so far today, and 514 new projects entered it. This week, 7,203 new projects and 100 new Hugging Face spaces were added. The top gainer today is rea with +14,624 stars.

Releases of the day

SGLang v0.5.22 adds support for new models and reflects a large contributor base. n8n's patch is limited to expression spans, which are now off by default. ComfyUI's patch addresses crashes tied to the MOSS image template workflow. For users, these are maintenance-level updates

Pricing moves

▼ 10%Qwen3 30B A3B Instruct 2507Alibaba · input $0.1 → $0.09 per 1M tokens
▼ 17%Kimi K3Moonshot AI · input $0.3 → $0.25 per 1M tokens
▼ 67%GLM 5.3 FlashZ.AI · input $0.15 → $0.05 per 1M tokens
▼ 25%DeepSeek V4.1 FlashDeepSeek · input $0.08 → $0.06 per 1M tokens
▼ 33%Nemotron 3.5 Super VLNVIDIA · input $0.15 → $0.1 per 1M tokens

Emerging projects

Research pick

Source: olud.ai tracking of 10,000+ open-source AI projects, 300+ models and live provider pricing. Figures are measured, not estimated. All releases · Live pricing · Latest in AI

← All AI news