AI News · November 3, 2026

Open-source AI gains 33,858 stars as two model prices jump

33,858 stars today86 new projects+2,242 dsh-Go

The open-source AI ecosystem added 33,858 GitHub stars today across 10,000+ tracked projects, with 86 new projects entering the index. Weekly totals reached 2,186 new projects and 100 new Hugging Face spaces. The top gainer was dsh-Go with +2,242 stars.

Two model prices rose sharply. OpenAI’s gpt-oss-120b input price increased 650% to $0.15 per 1M tokens. Google’s Gemma 4 26B A4B input price rose 14% to $0.08 per 1M tokens. These moves affect users of those specific models, though the broader ecosystem continues to expand.

Releases of the day

Seven open-source tools shipped minor updates today. OpenClaw v2026.10.10 improved backup safety by redacting credentials from Git backups. Ollama v0.40.14 fixed cloud model token limits for Claude Code. MinerU 4.1.5 normalized parse-path handling across CLI and Tasks API. Docling v2.146.0 added Unlimited-OCR 2 and LightOnOCR-3 presets. InvokeAI v6.14.4 fixed a startup crash caused by a Starlette library bug. Pydantic AI v2.65.0 added compatibility notes for agent tools. Windmill v1.858.0 introduced resolved runs on the home page.

Pricing moves

Two input price increases took effect today. OpenAI’s gpt-oss-120b rose 650% to $0.15 per 1M tokens. Google’s Gemma 4 26B A4B rose 14% to $0.08 per 1M tokens. No other price changes were reported.

▲ 650%gpt-oss-120bOpenAI · input $0.02 → $0.15 per 1M tokens
▲ 14%Gemma 4 26B A4BGoogle · input $0.07 → $0.08 per 1M tokens

Emerging projects

Three projects were detected today. elephant-goldfish-model (Python, 10,579 stars) provides open-source skills for designing and reviewing software with AI agents. mcd-budget-meal (HTML, 1,618 stars) is a McDonald’s MCP-based budget meal comparison tool. concord-mcp (TypeScript, 804 stars) offers live messaging for coding agents.

Research pick

A Hugging Face community paper, “Wave Forcing: Towards Speed-of-Light Streaming Video Generation,” received 79 upvotes. The authors propose a method for real-time long-video generation that avoids the efficiency and quality limits of chunk-wise autoregressive approaches.

The bigger picture

The sharp price increase on gpt-oss-120b and the steady influx of new open-source projects suggest a market adjusting to higher inference costs while the community continues to build. The research on streaming video generation indicates ongoing pressure for real-time capabilities.

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