Research library
Insights
Analysis of AI, automation, and software: what is changing, what holds up, and what it means for practical decisions.
173 articles · Page 1 of 15
Google Ships EmbeddingGemma 2: Text, Images, Video and Audio in One Vector Space
EmbeddingGemma 2 embeds text, code, image, video and audio in one 768-d space, ungated under Apache 2.0. Real sizes, self-reported scores, what to test.
Nano Banana 2.1: Half the Per-Image Price, 23 Days Left for Nano Banana 2 on the API
Google's Nano Banana 2.1 about halves the per-image price but triples input rates, and retires Nano Banana 2 on the Gemini API October 29. What to check.
Claude Sonnet 5.5: Compare Cost per Finished Task
Sonnet 5.5 keeps Sonnet 5 token prices. Its claimed savings come from using fewer tokens. How to test that claim in your workflow.
Holo4 Combines Computer Use and Tools, with Different Licenses
Holo4 spans GUI, code, MCP and API workflows. Its 27B and 35B-A3B checkpoints have different licenses. What to check before evaluation.
Reflection Announces Beam: 501B Open Weights Promised, Not Yet Released
Reflection announced Beam, a 501B/23B-active MoE with Apache 2.0 promised for later in October. No weights, license or report yet. What its tables show.
OpenAI DevDay 2026: Always-On Dots, GPT-6.1 Sol and Uneven Availability
OpenAI DevDay 2026 announced dots, GPT-6.1 Sol, Ultrafast, Agents API computer use and plugin extensions. What is live now, what it costs and what to test.
Perceptron Mk1.5 Adds Video Tracking for Embodied Agents
Perceptron Mk1.5 adds audio and video tracking. A September 28 assessment of its hosted API, context budget and integration constraints.
Claude Opus 5.5 Cuts Costs and Changes Conversation Handling
Anthropic's Opus 5.5 release lowers token prices. Its preserved-thinking rules also make conversation history a migration concern for custom agents.
GPT-6 Sol and Luna Bring Lower Prices to Agent Work
OpenAI releases GPT-6 Sol and Luna with lower prices and a shared million-token context window. Compare costs, caching and migration constraints.
Anthropic's Agent Metrics Put Review Delays in Focus
Anthropic proposes metrics for AI research and agent oversight. What coverage, review latency and escalation rates can tell an engineering team.
Bonsai 2 Makes a 27B Model Smaller. The Runtime Still Matters.
Bonsai 2 compresses a Qwen3.8-derived model into a 5.95 GB download. What the custom runtime and separate vision files mean for local use.
Grok 4.7: Check the Whole Agent Bill Before Switching
Grok 4.7 brings a 500K context window and configurable reasoning. What its API pricing, Fast availability and response behavior mean for migration.