The AI inference market is bifurcating into cheap commodity models and premium frontier systems, with GPT-4-class costs dropping from $20 to $0.40 per million tokens since 2022 while newer models like…
Google scrapped the base architecture for Gemini 3.5 Pro and rebuilt it from scratch, delaying the launch to a July 17 internal target after identifying performance gaps in mathematical reasoning, SVG…
A new study from the barkup-bench series reveals that LLM agents suffer from session drift when serializing tree state only once, with per-step success dropping from 98.8% to 83.8% for Claude Sonnet 4…
Researchers introduced CoC-Seduce, a benchmark for testing LLM rule adherence in semi-open textual environments, revealing that frontier models like GPT-5.4, Claude Sonnet 4.6, and Gemini 3.5 Flash ar…
Artificial Analysis launched AutomationBench-AA, an independent leaderboard for Zapier's AutomationBench, testing AI agents on 657 real SaaS workflow automation tasks across 40 simulated apps. Claude …
A developer tested RAG, fine-tuning, and long-context approaches on the same 1.1M-token knowledge base, finding that long-context cost 24x more than RAG and failed on mid-document facts, while fine-tu…
A developer demonstrated that six frontier large language models from OpenAI, Anthropic, and Google can be made to output zero bytes when prompted with 'Be silence.' and 'Be nothing.' using a specific…
Anthropic released Claude Sonnet 5 on June 30, 2026, but Roboflow evaluations show it ties with its predecessor Sonnet 4.6 on vision benchmarks at 70% accuracy, while being outperformed by Gemini 3.5 …
GitHub will deprecate Gemini 2.5 Pro and Gemini 3 Flash across all Copilot experiences on July 31, 2026, recommending Gemini 3.1 Pro and Gemini 3.5 Flash as alternatives. Users must update workflows a…
Amp rewrote its read_thread tool to handle long threads, which can now exceed 21 million tokens due to compaction. The new version uses a subagent with GLM 5.2 to extract information more accurately, …
Google DeepMind silently released Gemini 3.5 Flash via API on May 19, 2026. The model ranks #11 on LM Arena with a score of 1479.46 and #4 on agentic benchmarks with a score of 94, highlighting the ne…
Google announced a series of AI updates in June 2026, including the launch of Gemma 4 12B for local AI agents, computer use integration in Gemini 3.5 Flash, and new models Nano Banana 2 Lite and Gemin…
Google may be testing an upgraded Gemini Flash model on the LM Arena, with early impressions showing incremental quality improvements over the current version. The checkpoint, possibly labeled Gemini …
Alphabet's Google unveiled new AI tools and a redesigned search box at the I/O conference, powered by the Gemini 3.5 Flash model and new features Gemini Spark and Gemini Omni. The announcements are pa…
Anthropic released Claude Sonnet 5, a mid-tier model that delivers near-Opus-level performance in coding and agentic tasks at lower cost. The model outperforms previous Sonnet versions and competitors…
Google's Antigravity 2.0 platform enables developers to build multi-agent systems that orchestrate complex workflows by integrating with Model Context Protocol (MCP) and Google Cloud Data Services. Th…
Anthropic released Claude Sonnet 5, a cheaper and more agentic midsize model that can plan, use tools, and run autonomously. Priced at $2 per million input tokens through August, it offers performance…
NVIDIA released Nemotron 3 Ultra, a 550-billion-parameter open model under a fully permissive license, competitive with GPT-4.5 on code and reasoning. Z.AI launched GLM-5.2 with MIT-licensed weights, …
A new study finds that large language models from Google, Anthropic, and OpenAI exhibit gender-dependent triage disparities, giving young women significantly lower emergency room referral rates than m…
Five AI models, a lookup table, and a human football fan predicted 72 World Cup group-stage fixtures. Claude Sonnet 4.6 won with 67 points, while GPT-5.5 tied the human at 52 points, and the simple Ra…