Model wars: Anthropic Glasswing/Mythos cyber/visual, Claude vs ChatGPT switching/pricing, Gemini 3.1/Maps/UI/Nano/Gemma4/Grok/GLM-5.1/Qwen/Llama4/GPT-Image-2
Key Questions
What is Anthropic's Glasswing project?
Glasswing is a multimodal visual reasoning model with compositional anti-hallucinations and cyber capabilities like zero-day detection. It enhances AI's visual and security understanding. It's part of ongoing model improvements.
How has Claude's pricing and subscriptions changed?
Anthropic shifted Claude pricing, ending OpenClaw subscriptions amid high burn rates. This signals a new AI pricing era. Users are switching to ChatGPT, especially for coding.
What are the app rankings for major AI models?
OpenAI ranks #1, Gemini #2, Claude #3, and Grok #22 in app rankings. This reflects user adoption trends. ChatGPT maintains a lead over competitors.
What updates came for Gemini 3.1?
Gemini 3.1 Ultra features Maps integration with hallucinations, new UI on Android, and Nano/Gemma4 models. It includes AI-generated captions for Google Maps photos. Google refined the experience for better usability.
What achievements does GLM-5.1 have?
GLM-5.1 is #1 in open source and #3 globally on SWE-Bench Pro and Terminal-Bench. It's designed for long-horizon tasks. Available on Hugging Face.
Why are users switching between Claude and ChatGPT?
Users switch based on coding edges, pricing, and performance in 2026 comparisons. Claude excels in some areas but faces subscription shifts. ChatGPT holds strong overall.
What new features does Grok have?
Grok includes photo editing on X with AI tweaks and iPhone app availability. It's integrated into platforms like X. App rankings place it at #22.
How does Qwen3.5-VL perform?
Qwen3.5-VL is part of advancing open models like Llama4 and GPT-Image-2. It contributes to multimodal capabilities. Benchmarks and quants drive ongoing competitions.
Glasswing multimodal visual reasoning/compositional anti-hallucinations + cyber zero-days; Claude pricing shift (no OpenClaw subs, high burn)/vs ChatGPT switching (coding edge); Gemini 3.1 Ultra Maps hallucinations/UI/Android; App rankings OpenAI #1/Gemini #2/Claude #3/Grok #22; Gemma4 GGUF/GLM-5.1 #1 open/Qwen3.5-VL/Llama4. Ongoing TCO/quants/pricing/benches/safety/UI/apps.