Short, same-day coverage of notable AI model releases and announcements. These briefs give you the core facts — what was released, how it compares, and where to find more — without the padding.
Anthropic has launched Claude Fable 5.1, featuring a 1M-token context window, a 52.6% score on Terminal-Bench-Science, and a 75% reduction in prompt cache read costs to $0.25 per million tokens.
DeepSeek has officially open-sourced DeepSeek-V4-Flash-Vision-Exp, a 305-billion parameter native multimodal Mixture-of-Experts model published under the permissive MIT license on Hugging Face.
Tencent has open-sourced Hy4 Preview, a 770-billion parameter Mixture-of-Experts model activating 49B parameters per token with a 1M token context window on Hugging Face.
Gnani AI has launched Gnani Artha, a sovereign AI stack anchored by Evon 3.3, a 30-billion-parameter open-weight foundation model fine-tuned on Nvidia architecture for Indic enterprise workflows.
Singapore-based Sapiens AI has released Agnes 2.5 Pro Beta, a 1M-context proprietary reasoning model offering 159.5 tokens per second throughput and $0.10 per million input tokens.
Google released Gemini 3.8 Flash on September 2, 2026 — a 5.7-point jump over 3.7 Flash on autonomous software engineering at the same introductory price. A restricted Cyber variant, available only through the new Fairwind Program, targets government and critical infrastructure operators for agentic vulnerability remediation.