According to The Rundown AI, Meta released Muse Wonder, with Wonder Full scoring 56 on Artificial Analysis’ Intelligence Index and leading Opus 5.5 on DeepSWE, SWE-Marathon, and MemoryBench; it is available now via Meta Model API and Muse Code.

According to @bcherny, Claude Opus 5.5 built an interactive site and watercolor art for Acquired’s Home Depot episode.

According to TheRundownAI, Gemini 4 Argon leads 13 of 19 tests vs GPT6 Astra and Claude Opus 5.5, posts 1M-token output, and excels at long-context tasks.

According to @DeepLearningAI, Andrew Ng says early AI needs lightweight tests while mature products require rigor, plus updates on Claude Opus 5.5 metrics.

According to @godofprompt, Anthropic’s Opus 5.5 guide highlights 7 settings to boost quality, reduce latency, and improve tool use.

According to AnthropicAI, Claude Opus 5.5 matches Fable 5.1 on most tasks while costing 40% less than Opus 5.

According to karpathy, Claude Opus 5 used a 1M token run to generate 5,500 lines of Three.js code rendering LoTR, revealing promise and audit limits.

According to TheRundownAI, Kimi K3 tops Claude Opus 4.8 and rivals GPT 5.6 Sol and Fable at Sonnet pricing, signaling a strong open source push.

According to emollick, GDPval AA v2 relies on AI judges and unclear human ELO, limiting trust in Intelligence Index rankings.

According to @godofprompt, a one-shot prompt tested Claude Opus 4.8 vs Fable 5 on a full chess app with rules, AI levels, and animations.

According to @openclaw, the 2026.5.28 release adds Claude Opus 4.8, Krea via Fal, faster Gateway and plugins, plus Discord draft commentary.

According to @godofprompt, Anthropic says Opus 4.8 is 4x less likely to ship flawed code than 4.7, highlighting prompt engineering’s impact on reliability.

According to @AINewsOfficial_ on X, Gemini 3.5 Flash, Atlas updates, AGIBot X2 feat, and Claude Opus 4.8 performance claims are highlighted in a roundup.

According to @emollick, Claude Opus 4.8 wrote a paper from archived research while GPT5.5 Pro reviewed, finding one major error that Opus fixed.

According to @AnthropicAI, Claude Opus 4.8 improves judgment, transparency, and sustained autonomous work, and is available now at the same price.

According to @godofprompt, Opus 4.7 ignores project instructions and skips MCP configs; Anthropic acknowledged regressions versus 4.6 despite higher benchmarks.

According to @godofprompt, Claude Opus 4.7 follows instructions literally, lifts SWE-bench to 87.6% from 80.8%, and breaks 4.6-tuned prompts.

According to God of Prompt, DeepSeek V4 natively runs Claude Code and costs $0.14 per million tokens versus $5.00 for Claude Opus 4.7.

According to @SawyerMerritt, Anthropic raised Claude Opus API limits as xAI’s Colossus 1 adds over 300 MW capacity, expanding enterprise throughput.

According to @godofprompt, Genspark Build uses Claude Opus 4.7 to turn ideas into websites, apps, and code, enabling rapid product testing at startup speed.