AI agents can turn photos into 3D scenes, but still cannot reliably judge their own work
LEGO-Anything shows coding agents can build editable 3D scenes from photos, but LEGO-Bench exposes major gaps in accuracy and self-assessment.
Coding AI
Latest News and Analysis in Coding AI
Coding AI
Why coders may remain AI’s heaviest users
The Economist asks whether developers will outpace every other profession in AI use, highlighting coding’s unusual fit with measurable automation.
Goose challenges Claude Code’s $200 pricing with a local, open-source alternative
Block’s open-source Goose offers local AI coding agents without subscriptions, challenging Claude Code’s premium pricing while exposing trade-offs in quality and hardware.
DeepSeek reportedly assembles team to challenge Anthropic’s Claude Code
DeepSeek is reportedly forming a team to develop AI agents that could compete with Anthropic’s Claude Code, intensifying pressure in coding tools.
OpenAI-backed field report finds coding agents can overhaul research software, but not verify the science
An OpenAI-backed report says coding agents can speed research software upgrades dramatically, but experts still must verify scientific correctness.
OpenAI spotlights AI coding agents in scientific computing push, with genomics and legacy software modernization at the center
OpenAI says scientists are using AI coding agents to update legacy research software and speed genomics work, signaling a new enterprise AI battleground.
Anthropic unveils Claude Opus 5 with a coding focus, lower pricing and expanded API controls
Anthropic has introduced Claude Opus 5, pitching stronger coding performance, lower costs and new API features for agents and enterprise AI use.
Poolside releases Laguna S 2.1, betting better agent behavior can beat brute-force scale in coding models
Poolside released Laguna S 2.1, an open-weight coding model that it says rivals far larger systems by improving persistence and verification.
monday.com details how it runs production AI agents on Amazon Bedrock, signaling a more operational phase for enterprise coding agents
monday.com says it runs production AI agents on Amazon Bedrock, offering a rare look at the architecture and controls behind enterprise coding workflows.
OpenAI’s Codex is now hiding agent-to-agent instructions, raising transparency and reliability concerns for developers
OpenAI’s Codex now encrypts internal agent handoffs, limiting developer visibility into delegation and raising reliability questions for AI coding workflows.
OpenAI says roughly 30% of SWE-Bench Pro may be broken, challenging a key coding benchmark
OpenAI says about 30% of SWE-Bench Pro tasks may be broken, raising new doubts about how the AI industry measures coding models.
Japan’s labor crunch and legacy systems are turning AI coding agents into a serious enterprise pitch
Fortune reports Japan is emerging as a key market for Devin as aging software systems and a shrinking developer base sharpen demand for AI coding agents.
Alibaba's Qwen3.6-27B Beats Much Larger Models on Coding Benchmarks
Alibaba's open-source Qwen3.6-27B outperforms its 15x larger predecessor on most coding benchmarks with only 27 billion parameters.
Anthropic Releases Claude Sonnet 4.6 with Opus-Level Performance at Lower Cost
Anthropic launches Claude Sonnet 4.6, delivering frontier AI performance in coding, computer use, and agents with 1M token context window, just 12 days after Opus 4.6.
Anthropic's Claude Sonnet 5 Release Imminent with Major Performance Gains at Half the Cost
Industry reports suggest Claude Sonnet 5 will match Opus 4.5 capabilities while costing 50% less, with enhanced coding and agent capabilities.
Anthropic Launches Fast Mode for Claude Opus 4.6 with 2.5x Speed Boost
Anthropic introduces a fast mode for Claude Opus 4.6 that delivers responses up to 2.5 times faster, revolutionizing AI-powered software development and coding workflows.
Anthropic Releases Claude Opus 4.6: Most Capable AI Model with Enhanced Coding and Agent Teams
Anthropic launches Claude Opus 4.6, featuring improved coding skills, longer agentic task sustainability, and innovative agent teams functionality for enterprise applications.
Ads