AI News Roundup — October 14, 2026
A quieter day for new launches, so this roundup covers the genuinely verified, sourced items from the past week (October 6–14) — the current news cycle at time of writing. Reported factually, no speculation.
Anthropic opens a Singapore office and signs government MOU
Anthropic opened its fifth Asia-Pacific office, in Singapore, and signed a memorandum of understanding with Singapore's Ministry of Digital Development and Information. The agreement covers collaboration on AI model testing and security, research on AI adoption, and responsible use across government and industry.
Source: anthropic.com/news — "Anthropic opens Singapore office and signs MOU with the Government of Singapore" (Oct 14, 2026)
OpenAI rolls out GPT-6 with "Intelligent UI" in ChatGPT
OpenAI began rolling out GPT-6, a new model family for ChatGPT, together with "Intelligent UI" — responses that can include interactive elements such as graphics, charts, forms and in-conversation tools. The rollout started with ChatGPT Plus, Pro, Business and Enterprise tiers, with Free and Go tiers following a day later (Enterprise access depends on workspace admin settings).
Source: openai.com — "GPT-6 and Intelligent UI for everyone" (Oct 7, 2026)
OpenAI publishes a safety update for GPT-6 Sol and GPT-6 Luna
Alongside the GPT-6 rollout, OpenAI published an October 2026 deployment-safety update covering its GPT-6 Sol and GPT-6 Luna model variants.
Source: deploymentsafety.openai.com — "GPT-6 Sol and GPT-6 Luna: October 2026 update" (Oct 7, 2026)
Anthropic ships Claude Haiku 5.5
Anthropic released Claude Haiku 5.5, the latest entry in its fast/lower-cost Haiku model line, as its newest featured model announcement of the month.
Source: anthropic.com — "Introducing Claude Haiku 5.5" (Oct 7, 2026)
Mistral previews Mistral Large 4
Mistral AI opened a public preview of Mistral Large 4 (internally "ML4"), a roughly 1-trillion-parameter multimodal model with about 52 billion active parameters, trained in Mistral's European datacenters and available via the Mistral Studio API. Mistral says open weights are planned to follow later in the month, and that the model leads among open-weight models on cybersecurity, finance and legal benchmarks it tested.
Source: mistral.ai/news — "Mistral Large 4" (Oct 6, 2026)
Meta ships its first frontier model, Muse Wonder
Meta introduced Muse Wonder and Muse Wonder Full, described as its first frontier-class models, aimed at agentic tasks and available through Meta's Model API and Muse Code (Wonder Full is limited to Muse Code subscribers). Meta's own performance claims for the models have not been independently verified by ModelScore at time of writing.
Source: letsdatascience.com — "Meta Launches Muse Wonder and Muse Wonder Full for Agentic Tasks" (reporting on Meta's official announcement, Oct 12, 2026)
Methodology: items are drawn from official vendor newsrooms where available, cross-checked against independent reporting. We link the primary source for every item and do not editorialize on capability claims we haven't verified ourselves. On a thin news day, we report that plainly rather than filling space.