AI News Feed
Market watch
Category

Large Language Models

Latest news, developments and reporting on Large Language Models.

Large Language Models · Products & Applications · Research

Kyutai Releases Voice of Reason, Speech-Native Math Models With 77.1% Spoken GSM8K Accuracy

Kyutai released two open-weight speech-to-speech models that solve spoken math, reaching 77.1% on GSM8K after RL.

Large Language Models · Companies · Products & Applications

Xiaomi Open-Sources MiMo-V2.6 as Alibaba Expands Qwen Agent Platform and Jev Offers 120 Million Tokens

On September 22-23, 2026, Xiaomi open-sourced MiMo-V2.6 with 1M-token context and new RL post-training, Alibaba upgraded its Qwen AI platform with Agent Studio, and Jev issued 120 million tokens to approved users.

Large Language Models · Companies

OpenAI Releases GPT-6 Sol and Luna as Anthropic Launches Claude Opus 5.5

OpenAI launched cheaper GPT-6 Sol and Luna while Anthropic released Claude Opus 5.5, intensifying competition on price, coding and agent performance.

Large Language Models · Companies · Products & Applications

OpenAI Cuts GPT-6 Prices as Anthropic Releases Claude Opus 5.5

OpenAI launched GPT-6 Sol and GPT-6 Luna with API prices about half the previous generation, while Anthropic released Claude Opus 5.5 at lower cost and faster speed. The same-day moves push frontier-model competition toward cost per task and long-running agents.

AI Chips & Compute · Large Language Models · Companies

DeepSeek to Brief UN Security Council on AI Risks; Alibaba Says Qwen4 in Training

DeepSeek and other Chinese AI firms have been invited to brief the UN Security Council on AI risks. Alibaba says Qwen4 is in training and a new AI chip is due in early 2027, while Huawei holds back a global AI chip push as domestic demand outstrips supply.

Large Language Models · AI Chips & Compute · Products & Applications

Anthropic and OpenAI Release New Models as Qualcomm, OPPO and Beats Unveil Hardware

On Sept. 22–23, 2026, Anthropic and OpenAI released new AI models, Qualcomm announced two Snapdragon 8 platforms, and OPPO and Beats launched new devices, according to Shaoshupai.

Large Language Models · Products & Applications · Companies

OpenAI and Anthropic Cut Model Costs as AI Expands Into Phones

OpenAI and Anthropic released cheaper frontier models on Sept. 22, while Apple, OPPO, Qualcomm and Honor pushed new AI capabilities into devices and Meta's Muse agent faced a macOS security issue.

Cybersecurity · Large Language Models

Cisco Talos Details Windows Malware That Lets Four LLMs Vote on Its Next Move

Cisco Talos released an open-source toolkit for hunting AI-enabled malware and detailed CLOSEDQUORUM, a Windows credential stealer with no command-and-control server that puts each tactical decision to a majority vote of four commercial language models.

Large Language Models · Companies · Products & Applications

Anthropic Releases Claude Opus 5.5 as OpenAI Counters With Cheaper GPT-6 Sol and Luna

Anthropic released Claude Opus 5.5 on Sept. 22 with a 20% price cut, and OpenAI minutes later launched GPT-6 Sol and Luna at half the cost of their predecessors, according to SiliconANGLE.

Large Language Models · Products & Applications

Anthropic launches Claude Opus 5.5 with faster output and about 40% lower running costs

Anthropic released Claude Opus 5.5 on Sept. 22, promising faster, cheaper and more accurate output than Opus 5, along with stronger safety safeguards.

Large Language Models · Companies · Products & Applications

Anthropic Releases Claude Opus 5.5 as OpenAI Unveils GPT-6 Sol and Luna

Anthropic released Claude Opus 5.5 and OpenAI introduced GPT-6 Sol and Luna on Sept. 22, 2026, offering stronger performance and lower prices as debate continues over pacing frontier AI.

Products & Applications · Companies · Large Language Models

Anthropic and OpenAI Release Cheaper Models in First Launches Since Slowdown Call

Anthropic and OpenAI on Tuesday released lower-cost AI models, the first launches since Anthropic CEO Dario Amodei called for an industry-wide slowdown on advanced AI development.

Large Language Models · Companies

Anthropic releases Claude Opus 5.5 with Fable-level performance, lower prices and stricter safeguards

Anthropic released Claude Opus 5.5 on September 22, 2026, saying it matches Fable 5.1 on most work, costs about 40% less to run and adds safeguards that reroute cybersecurity and biology requests.

Large Language Models · Products & Applications · AI Chips & Compute

Alibaba Launches Enterprise Agent Infrastructure and Multimodal Models; Intel Says CPU Value Is Rising

On Sept. 22, 2026, Alibaba's Qwen Office released Enterprise Context and QwenNote A2 for enterprise agents, while Alibaba Cloud unveiled a full multimodal model lineup. Intel separately said AI agents are shifting data-center compute ratios toward one CPU per GPU.

Policy & Regulation · Cybersecurity · Large Language Models

Altman, Amodei Expected to Address UN Security Council AI Meeting

OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei are expected to address the U.N. Security Council on Wednesday in New York, a meeting on artificial intelligence that comes as safety concerns mount and President Trump vows to encourage the technology.

Large Language Models

OpenAI releases GPT-6 Sol and Luna, claiming lower cost and fewer mistakes

OpenAI released GPT-6 Sol and Luna, expanding its GPT-6 family after Astra. The models cost half as much via API as GPT-5.6 versions and roll out across ChatGPT Work, Codex, and the API, with Luna available to Free and Go desktop users.

Products & Applications · Large Language Models

Vibe Coding Expands Into a Software Category, but First Websites Still Need Refinement

TechRadar's guide says vibe coding has become its own software category, but building a first website still requires refining AI output, choosing hosting, and troubleshooting. It outlines tool types and initial steps.

Companies · Large Language Models

AstroForge to Put AI in Command of Its Next Spacecraft

AstroForge will fly Solo, an in-house transformer AI, on a 2027 spacecraft after two missions suffered anomalies.

Large Language Models · Products & Applications

Xiaomi open-sources MiMo-V2.6 omnimodal AI model family

Xiaomi has released and open-sourced the MiMo-V2.6 series, led by the flagship Pro model, with agentic benchmark scores near leading closed systems and API pricing below comparable overseas flagships.

Companies · Large Language Models

Mistral acquires Paris-based adtech startup Pimento in €12.7m cash-and-shares deal

Mistral acquires Paris-based adtech startup Pimento in a €12.7m cash-and-shares deal, its third acquisition this year, Sifted reports.

Large Language Models · Companies

Four Chinese Models Pass 10 Trillion Weekly Tokens as DeepSeek and Hunyuan Take Near Half of Usage

Four Chinese models each exceeded 10 trillion tokens in the week to Sept. 18, with DeepSeek's two Flash models and Tencent's Hy4 preview close to half of domestic usage, the AGI Market Watch says.

Products & Applications · Large Language Models

Windows Task Manager Creator Ships TMOG, a Vibe-Coded Cross-Platform Rival Built From a 107-Page Spec

Dave Plummer, who wrote the original Windows Task Manager in 1996, has used Claude Code and a 107-page spec sheet to build TMOG, a cross-platform task manager now in beta for macOS, Windows and Linux, with a $39.95 Pro tier.

Companies · Large Language Models

WebArena Author Shuyan Zhou Joins Meta Superintelligence Labs to Build an AI Browser

Shuyan Zhou, the researcher behind the WebArena benchmark, says she left academia early this year to join Meta's Superintelligence Labs, where her main focus is building an AI browser.

Companies · Large Language Models

Xiaomi's MiMo-V2.6 tops open-source ranking after six-day, $3.5M live RL run

Xiaomi completed a six-day live-streamed reinforcement learning run for MiMo-V2.6, spending about $3.5 million. The Pro model scored 46 on the Artificial Analysis Intelligence Index, ranking first among open-source models.