Anthropic's Fable 5 Adoption Slows, Hinting at Corporate AI Spending Ceiling
Anthropic's Fable 5 captured 11.4 percent of token spending on its models in its first month, compared to 23 percent for OpenAI's GPT-5.6 Sol.
Model Release
428 articles
New AI models as they ship — large language models, multimodal systems, and open-weight releases. Capabilities, context windows, pricing, and availability, summarized so you can compare what is actually new.
LATEST NEWS
-
-
Hugging Face Launches Luth-2 Models for Enhanced French Language Capabilities
Hugging Face introduces Luth-2-0.8B and Luth-2-2B, two French language models that set new benchmarks in math, code, and instruction following, outperforming all models in their size class.
-
SpaceXAI's Grok 4.6 Ties OpenAI's Best Model on Benchmark
SpaceXAI's Grok 4.6 scores 61 on the AA Index, matching OpenAI's GPT-5.6 Sol, according to the Artificial Analysis Intelligence Index.
-
Cohere Launches North-Micro-Vision-Instruct, 2.4B-Parameter Vision-Language Model
Cohere released North-Micro-Vision-Instruct, a 2.4B-parameter open-weight vision-language model, under Apache 2.0 license on August 12, 2026.
-
Google Unveils Pixel 11 Lineup With Gemini Features
Google announced the Pixel 11 series at its Made by Google 2026 event, including a new tracking tag and AI-powered updates.
-
DeepMind Launches SL2T Model for Sign Language AI
DeepMind introduces SL2T, a sign-language-to-text model, enabling Deaf users to sign to their phones in 2026. The model supports ASL to English and is available in Gboard and Live Transcribe on Pixel 11.
-
Microsoft's MAI Code 1.1 Flash Falls Short Against Deepseek
Microsoft's MAI Code 1.1 Flash model underperforms Deepseek-V4-Flash-0731 on both price and performance, according to benchmarks.
-
Nvidia's Nemotron 4 Aims for One Trillion Parameters
Nvidia's Nemotron 4, set for release this fall, will have at least one trillion parameters, a scale already surpassed by Chinese labs.
-
HuggingFace Model Creator Blocks Critics for Asking About Benchmarks
User banned after questioning benchmark scores of Qwen models, highlighting a pattern of suppression on HuggingFace.
-
AWS and OpenAI Launch Cyber Defense Models on Amazon Bedrock
AWS and OpenAI have made Daybreak Red and Daybreak Blue available on Amazon Bedrock, offering advanced cybersecurity tools to eligible customers.
-
ONESTRUCTION Builds Ishigaki-IDS Foundation Model with AWS GenAIIC
ONESTRUCTION, Inc. developed the Ishigaki-IDS foundation model, designed for construction BIM workflows, using AWS GenAIIC technical advisory.
-
Nvidia's Nemotron 3.5 Lightning Prioritizes Speed Over Intelligence
Nvidia's Nemotron 3.5 Lightning matches OpenAI's gpt-oss-120b on intelligence benchmarks with a quarter of the parameters and delivers the fastest inference speeds in its class.