IBM Think 2026 closed on May 7 with a sweep of final product announcements. Meanwhile, a new agent benchmark running on real production websites published results that put Claude Sonnet 4.6 at the top, and Anthropic's $1.5 billion joint venture structure became fully confirmed.
IBM Think 2026 — Final Day Highlights
IBM's annual Think conference concluded its four-day run in Boston on May 7 with formal general availability announcements for several previewed products. IBM Sovereign Core — the platform that embeds governance policy at the infrastructure runtime level for regulated, cross-border environments — reached GA. IBM Bob (Pro, Pro+, Ultra, and Enterprise SaaS) launched as an end-to-end software development partner covering code generation, testing, security, and deployment across the full SDLC. Unlike point-in-time coding assistants, Bob operates across the entire application lifecycle. The next generation of IBM watsonx Orchestrate for multi-agent orchestration also received its full release, enabling enterprises to build, deploy, and manage thousands of agents built by different teams across an organisation.
IBM Docling for watsonx — a document intelligence platform that converts documents into structured Markdown, JSON, and HTML for RAG workflows — and OpenRAG on watsonx.data, an open agentic retrieval framework, both shipped alongside the conference close.
Claude Sonnet 4.6 Tops ClawBench — the First Real-Website Agent Benchmark
Researchers from UBC and Vector Institute published ClawBench, a new evaluation framework of 153 tasks across 144 live production websites in 15 categories — including completing purchases, booking appointments, and submitting job applications. Unlike prior benchmarks that ran in sandboxes, ClawBench operates on real production sites, intercepting only the final submission request to keep evaluation safe. Claude Sonnet 4.6 achieved the top score of 33.3% among all frontier models tested. The benchmark captures five layers of behavioural data per run: session replays, screenshots, HTTP traffic, agent reasoning traces, and browser actions — scored by an agentic evaluator that produces step-level diagnostics.
Anthropic $1.5B Joint Venture — Structure Confirmed
The full structure of Anthropic's private equity joint venture is now confirmed. Anthropic, Blackstone, and Hellman & Friedman each contributed approximately $300 million; Goldman Sachs contributed $150 million. Apollo Global Management, General Atlantic, Leonard Green, GIC, and Sequoia Capital also participated. The vehicle operates as a forward-deployed enterprise services firm — embedding Claude directly into the operations of PE-backed portfolio companies. CFO Krishna Rao said the structure exists because enterprise demand for Claude is "significantly outpacing any single delivery model."
Catch up on the full week's AI news at the May 2026 AI News Hub.