AI News · October 24, 2026

DeepSeek V4 Pro and Qwen models see sharp price increases

45,401 stars today65 new projects+7,121 rea
By olud.ai editorial · built from our own tracking data, published daily

DeepSeek V4 Pro 0423 input price rose 61% to $0.71 per 1M tokens. Three Qwen models from Alibaba also increased, with Qwen3.8 27B up 400% to $0.2 per 1M tokens. GLM 5.3 Flash from Z.AI saw the largest jump at 1400% to $0.15 per 1M tokens.

The ecosystem gained 45,401 GitHub stars across 10,000+ tracked projects today. 65 new projects entered the index. This week 1,972 new projects and 100 new Hugging Face spaces were added. The top gainer "rea" added 7,121 stars.

Releases of the day

Five minor releases shipped yesterday. Paperclip v2026.1023.0 carries 200 commits and was promoted from beta. CrewAI 1.15.30 fixes AGENTS.md login evaluation and stops after three rounds. LangGraph 1.2.17 applies an explicit recursion limit equal to the default. Pydantic AI v2.61.0 warns when provider-specific thinking settings override broader settings. Windmill v1.848.2 tightens file handling in deno and bun for executor isolation.

Paperclip v2026.1023.0 Paperclip v2026.1023.0 · > Released: 2026-10-23 · Paperclip v2026.1023.0 carries 200 commits, promoted from 2026.1012.1-beta.0. The headline is private work: a task or project... CrewAI 1.15.30 What's Changed · Bug Fixes - Fix AGENTS.md to start the login process while evaluating and stop after three rounds. · Documentation - Update snapshot and changelog for... LangGraph 1.2.17 Changes since 1.2.16 · * release(langgraph): 1.2.17 (#9418) * fix(langgraph): apply an explicit recursion_limit equal to the default (#9417) * fix(langgraph): keep the caller's... Pydantic AI v2.61.0 What's Changed ⚠️ Compatibility Notes * Warn when a provider-specific thinking setting from a broader scope overrides thinking by @dsfaccini in... Windmill v1.848.2 1.848.2 (2026-10-24) · Bug Fixes · * tighten file handling in deno/bun, stricter codebase tarball,executor isolation (#11990) (ccf8892) · --- This PR was generated with Release...

Pricing moves

Seven models changed input prices. Three increased: DeepSeek V4 Pro 0423 up 61%, Qwen3.5-9B up 113%, Qwen3.5-122B-A10B up 54%, Qwen3.8 27B up 400%, GLM 5.3 Flash up 1400%. Three decreased: Gemma 4 26B A4B down 12%, Mistral Small 3.2 24B down 11%, GLM 5.3 down 89%.

▲ 61%DeepSeek V4 Pro 0423DeepSeek · input $0.44 → $0.71 per 1M tokens
▼ 12%Gemma 4 26B A4BGoogle · input $0.08 → $0.07 per 1M tokens
▼ 11%Mistral Small 3.2 24BMistral AI · input $0.09 → $0.08 per 1M tokens
▲ 113%Qwen3.5-9BAlibaba · input $0.08 → $0.17 per 1M tokens
▲ 54%Qwen3.5-122B-A10BAlibaba · input $0.26 → $0.4 per 1M tokens
▲ 400%Qwen3.8 27BAlibaba · input $0.04 → $0.2 per 1M tokens
▼ 89%GLM 5.3Z.AI · input $0.7 → $0.08 per 1M tokens
▲ 1400%GLM 5.3 FlashZ.AI · input $0.01 → $0.15 per 1M tokens

Research pick

WeVisOne-VL from WeChat unifies object-level and pixel-level visual perception. The paper claims existing vision-language models remain unreliable for fine-grained visual tasks. The work received 229 upvotes on Hugging Face.

WeVisOne-VL: Unified Visual Perception from Objects to PixelsTianming Liang, Junyi Pan, Yuxi Li et al. · 229 community upvotes · Existing vision-language models (VLMs) remain unreliable for fine-grained visual perception. In this report, we present WeVisOne-VL (WeChat Vision All-in-One), a powerful VLM that unifies object-level and pixel-level perception through stru

The bigger picture

Alibaba and DeepSeek raising prices on multiple models signals a shift toward monetization in the open-weight space. The steep GLM 5.3 Flash increase suggests providers are recalibrating pricing tiers.

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