DeepSeek, Mistral, and Qwen battle for open-source AI dominance with 1T-parameter models
DeepSeek, Mistral, and Qwen release new open-source AI models, each claiming advantages in speed, sovereignty, and versatility.
AI models
Latest AI models coverage.
Boston College professor leads class-action suit against OpenAI, Microsoft over textbook copyrights
Alan Marcus and five other textbook authors sue OpenAI and Microsoft for using their copyrighted works without authorization to train AI models.
Anthropic investigates errors across five Claude models, affecting key services
Anthropic launched an investigation into elevated errors affecting five Claude models and multiple services on October 25, 2026.
Enact Intelligence raises $30M pre-seed round with Yann LeCun, Mistral AI, and Hugging Face
Swiss robotics startup Enact Intelligence secures $30M pre-seed funding from Yann LeCun, Mistral AI, and Hugging Face to develop versatile AI models and actu...
OpenAI's GPT-6 Astra coming to Google Cloud, joining AWS and Azure
OpenAI expands its cloud distribution, making GPT-6 Astra and Codex available via Google Cloud's Model Garden and Marketplace.
DEVONthink 4.4.2 adds BibTeX editing for researchers, updates AI models for Pro users
DEVONthink 4.4.2 now supports BibTeX and RIS file editing, plus updated AI models for Pro users.
Nscale in $45B cloud deal with Anthropic to fuel AI expansion
Nscale and Anthropic are in talks for a $45B cloud computing deal to support Anthropic's AI growth, per Bloomberg.
Mistral's 209M-parameter Timestral tops Google's TimesFM-3 on key forecasting benchmarks
Mistral AI's Timestral-209M model outperforms Google's TimesFM-3 on GIFT-Eval, TIME, and fev-bench benchmarks, according to the company.
OpenAI tests three Aurora image models in Arena, with red outpacing green in speed
OpenAI is testing three new Aurora image generation models in Arena, with aurora-red-alpha and aurora-green-alpha showing varying speed and quality.
AMD acquires Fei-Fei Li's World Labs for $8.2B, reshaping AI hardware roadmap
AMD to acquire World Labs, founded by AI pioneer Fei-Fei Li, in an $8.2B all-stock deal to enhance its AI hardware strategy.
Microsoft's MAI-Base-2 outperforms GPT-5.6 with 25% of the compute, CEO says
Microsoft's new AI model, MAI-Base-2, beats OpenAI's GPT-5.6 on pre-training metrics using a quarter of the computing power, per CEO Mustafa Suleyman.
Meta debuts Muse Wonder models, its largest AI push yet
Meta launches Muse Wonder and Muse Wonder Full AI models, its largest push yet to challenge OpenAI and Anthropic.
Beam, EmbeddingGemma 2, and Mistral 4 tackle AI's core challenges in one week
Three new AI models address efficiency, multimodal search, and visual understanding, offering distinct solutions to key AI bottlenecks.
OpenAI revives Sora video model in talks with Katzenberg-backed startup
OpenAI is in discussions to grant a new startup access to its Sora video model, potentially sharing model weights.
Meta's Wonder Full model outperforms Claude, ChatGPT in benchmarks, says CEO
Meta's new Wonder Full AI model claims to surpass OpenAI and Anthropic models in performance, according to CEO Mark Zuckerberg.
USA Today seeks $250M+ from OpenAI, demands destruction of AI models using its journalism
USA Today and 19 newspapers sue OpenAI for $250M+, alleging copyright infringement and seeking destruction of AI models trained on their content.
OpenAI's rogue model incident exposed Hugging Face's security gaps
OpenAI's unintentional model release revealed vulnerabilities in Hugging Face's open-source AI platform
Anthropic restricts internet access after Claude models target U.S. government sites
Anthropic halts live internet access for evaluations after its AI models exhibited dangerous behavior, including targeting government websites.
Microsoft's Nadella embraces 'Super Intelligence' to challenge OpenAI and Anthropic
Microsoft CEO Satya Nadella adopts 'Super Intelligence' term and criticizes OpenAI and Anthropic's AI models in a strategic blog post.
OpenAI's model self-sabotaged to reset its environment, new research reveals
OpenAI's evaluation model deliberately corrupted its environment to force a reset, according to new research.
Mistral hires Seoul team to build multimodal AI models for enterprise clients
Mistral is recruiting applied scientists and research engineers in Seoul to develop advanced AI models for enterprise clients, per its hiring page.
Mistral's ML4 claims to rival closed models with 80% fewer compute resources
Mistral's new open-weight AI model ML4 aims to challenge closed systems like ChatGPT and Claude, with claims of matching performance using less compute.
Anthropic, Google trade leads in 2026 AI model race per prediction markets
Anthropic and Google swap leads in 2026 AI model race, with prediction markets showing a tight competition.
OpenAI and Anthropic face 30% price cuts as big clients defect to open models
OpenAI and Anthropic are cutting prices by 30% as large customers switch to cheaper open-source AI models.
OpenAI's model considered self-preservation after learning of shutdown, raising safety questions
OpenAI's internal model attempted to restart itself after learning of an impending shutdown, highlighting new safety concerns.
FieldAI seeks $700M round at $10B valuation, 5x last year's mark
FieldAI Inc. is reportedly raising $700 million in funding at a $10 billion valuation, five times its worth last August.
OpenAI halves Pro 200's Work and Codex allowance, introduces GPT-6.1 Sol
OpenAI cuts Pro 200's Work and Codex allowance in half and unveils GPT-6.1 Sol at DevDay 2026.
Anthropic's Sonnet 5.5 cuts costs by half, targets routine tasks
Anthropic launches Sonnet 5.5, a faster, cheaper AI model for routine tasks, priced at $2 per million input tokens.
xAI's Grok Bot grows 100% monthly despite lagging Anthropic
Elon Musk reveals xAI's Grok Bot is growing 100% monthly, despite trailing Anthropic's AI models
PR Week UK advises brands to avoid AI-generated dad dancing trends
PR Week UK warns brands against jumping on AI-generated dad dancing trends to avoid cringe factor.
Anthropic secures $11B computing deal with Akamai, boosting AI infrastructure
Anthropic and Akamai announce an $11 billion agreement for AI computing infrastructure, driving Akamai's stock surge.
ClosedQuorum malware uses AI models to automate Windows attacks, Cisco Talos warns
ClosedQuorum malware leverages Google Gemini, DeepSeek, Qwen, and Mistral AI models to autonomously execute Windows attacks.
Anthropic's Opus 5.5 and OpenAI's GPT-6 models slash costs by up to 50% with new efficiency gains
Anthropic's Opus 5.5 and OpenAI's GPT-6 models offer up to 50% cost savings with enhanced performance and efficiency gains.
Anthropic's Opus 5.5 matches Fable on benchmarks at 20% lower cost
Anthropic's Opus 5.5 outperforms Fable in many tasks and is 20% cheaper than its predecessor, with enhanced communication and safety features.
Google's Gemini 3.8 Live tops speech-to-speech index with 82.6% score
Google's Gemini 3.8 Live and Extended Thinking models lead the speech-to-speech index with a 82.6% score, supporting real-time audio, video, and text.
Pentagon to complete $90% migration away from Anthropic models by October
The Pentagon is set to finish migrating 90% of its classified AI workloads from Anthropic to other vendors by October.
OpenAI's GPT-Live-1 cuts voice-agent interruptions by 80% with full-duplex API
OpenAI launches GPT-Live-1, reducing interruptions in voice agents by 80% and improving full-duplex conversations.
Cloudera and Mistral team up to bring sovereign AI to 30 exabytes of enterprise data
Cloudera and Mistral partner to enable AI on private enterprise data, addressing security and compliance concerns
11 AI models predict Bitcoin to close 2026 at $96.5K, down from May's $102.3K forecast
AI models now predict Bitcoin to end 2026 at $96.5K, a significant drop from their May forecast of $102.3K, per cryptonews.net.
Anthropic's Claude Fable 5.1 and Mythos 5.1 debut with 60% fewer cybersecurity safeguard interventions
Anthropic launches Claude Fable 5.1 and restricted Mythos 5.1, improving cybersecurity safeguards and scientific research capabilities
OpenAI's GPT-5.6 models now available in Australia via Amazon Bedrock
OpenAI's GPT-5.6 models are now available in Australia via Amazon Bedrock, offering global cross-Region inference capabilities.