Easy AI Daily | November 20, 2025
A roundup of AI industry news for November 20, 2025, compiled by the Easy AI teaching project.
Model Updates & Releases
Google Releases Gemini 3 Pro Image (Nano Banana Pro)
- Supports Google Search grounding, 2-4K resolution, text-in-image generation/editing
- Pricing: $0.134 per 2K image, $0.24 per 4K image
- Available in Gemini App/API, LM Arena, Hugging Face Spaces, Together AI
- Early demos show accurate infographics and chart annotation; text rendering error rate dropped from 56% to 8%
- Links: Pricing | Announcement | LM Arena | Hugging Face Spaces | Together AI | Error rate data | SynthID watermark
- Fully open source (Apache-2.0), including code, data, training recipes, checkpoints
- 32B Think variant with long chain-of-thought for complex reasoning
- Architecture retains post-norm; 7B uses sliding-window attention for KV cache, 32B uses GQA
- RL infrastructure gives 4x faster experiments; emphasis on decontaminated evaluation (e.g., random reward tests)
- Links: Announcement reaction | Architecture analysis | HF collection
- SAM3 unifies image/video segmentation with text/visual prompts, 2x performance, 30ms inference
- SAM3D enables 3D reconstruction from a single image
- Data engine: 4M phrases, 52M masks; open source with commercial use allowed
- Links: SAM3 | SAM3D | GitHub
- Designed for long-horizon, detail-heavy tasks; first native multi-context-window support via compaction
- SOTA on SWEBench; available only through ChatGPT plans, no API
- Links: Release blog | Announcement
- Deep Cogito's Cogito 2.1 ranks #18 overall, top 10 among open models; hosted on Together and Fireworks
- Links: Model | Leaderboard
- OpenAI shares 13 early experiments showing GPT-5.1 accelerating math, physics, biology, and materials science research; 4 helped solve previously unsolved problems. Links: Overview | arXiv paper
- Perplexity Comet browser: launches on Android, Mac, Windows; voice-first browsing, supports Kimi-K2 Thinking and Gemini 3 Pro; Pro/Max users can create slides, tables, documents. Android release
- Cursor beta debug mode: log ingest server auto-instruments code; agents verify hypotheses from logs rather than guessing
- MemMachine Playground: open Hugging Face Space supporting GPT-5, Claude 4.5, Gemini 3 Pro with persistent AI memory. Playground
- DSPy Proxy: new repo aryaminus/dspy-proxy built from a single prompt to simplify DSPy agent development. Repo
- A user built a 6-node NVIDIA Jetson Spark cluster for NCCL development and prototyping pre-B300 workflows. Reddit
- GPU MODE discussions: GEMM optimization, texture vs constant cache, AMD MI300X DMA collectives (+16% on large transfers), BF16 conversion gaps in TensorRT. GEMM blog | DMA paper
- Mojo 0.25.7 nightly shows a major regression: llama2.mojo throughput dropped from ~1000 tok/sec to ~170 tok/sec on Mac M1
- Community reports jailbreaks against Gemini 3 Pro, Grok (shell access), and Claude 4.5 (trust-building bypass)
- SynthID watermark bypass found via reve-edit "do nothing" prompts, or by asking the model directly whether content is AI-generated
- LMArena: debates on Nano Banana Pro quality, SynthID bypass, GPT-5.1 vs Gemini 3 Pro
- Perplexity: Gemini 3 Pro coding praised over Claude Sonnet 4.5; Comet RAM usage concerns
- Unsloth AI: RAM price surge (64GB now ~$400)
- Eleuther AI: KNN vs quadratic attention, RE-Bench results claimed superhuman by IntologyAI
- Yannick Kilcher: AI CEO benchmark (LLMs worse than humans at long-horizon planning), NVIDIA Q3 earnings
- modelcontextprotocol.io migrates from Anthropic to community control
- OpenRouter reports 500 errors and agentic LLM mid-run pauses; Grok 4.1 free until Dec 3
AI2 Releases Olmo 3 Open LLM
Meta Releases SAM3 and SAM3D
OpenAI Launches GPT-5.1 Codex Max
Cogito 2.1 Enters WebDev Arena
Research & Science
Tools & Platforms
Hardware & GPU
Safety & Jailbreaking
Community Highlights
Other News
*Source: Easy AI teaching project*