Series
AI Model Notes
Observations on AI models, open-source projects, and inference tools, with a focus on positioning, capability limits, deployment needs, and practical model selection.
3 posts · Read in order
- 1
baidu/Unlimited-OCR: A New Option for Long-Document OCR
baidu/Unlimited-OCR is Baidu's open-source long-document OCR and document parsing model, focused on converting PDFs, papers, reports, and books into Markdown. Its strengths are long-context parsing, relatively friendly GPU requirements, and open-source usability, but it is better suited to engineering workflows than simple screenshot OCR.
June 30, 2026 · 7 min read
- 2
Claude Sonnet 5: A More Agentic Sonnet Model
Claude Sonnet 5 is Anthropic's next-generation Sonnet model, with stronger reasoning, tool use, coding, and long-horizon task execution. It narrows the gap with Opus 4.8 in agentic scenarios while covering more everyday development and knowledge-work tasks at a lower price.
July 1, 2026 · 7 min read
- 3
GPT-5.6 Sol, Terra, and Luna: OpenAI Splits Agentic Coding Into Three Tiers
OpenAI previewed the GPT-5.6 family: Sol is the frontier tier, Terra is the balanced everyday tier, and Luna is the fast low-cost tier. The real shift is that model selection moves from 'which model is strongest' to 'which tier should this task be routed to'.
July 10, 2026 · 7 min read