↗TiaoTechIdeas in motion
中文
☰
Home/Stay curious. Stay in the loop.
TIAOTECH / NEWS

Stay curious. Stay in the loop.

Selected industry news from AIHOT. Follow advances in AI, products and technology.

AIHOT · Selected summaries with original sources · Nov 5, 2026, 9:00 AM

AI models · Hugging Face Blog

Hugging Face and Artificial Analysis Release Text-to-Image Model Leaderboard and Arena ↗

Hugging Face and Artificial Analysis jointly released the Text to Image Leaderboard and Arena, calculating ELO scores based on over 45,000 human image preference data points. The evaluation covers mainstream open-source and closed-source text-to-image models such as Midjourney, DALL·E, Stable Diffusion, and Playground; users can participate in voting and receive personalized model rankings.

Industry · Hugging Face Blog

Hugging Face Discloses Unauthorized Access to Spaces Platform Secrets ↗

Hugging Face officially disclosed a recent unauthorized access incident on its Spaces platform, where some Spaces secrets may have been accessed. The company has revoked relevant HF tokens, encouraged users to switch to fine-grained access tokens, removed org tokens from infrastructure, integrated KMS, enhanced secret leak detection and automatic invalidation capabilities, and reported the incident to law enforcement and data protection authorities.

AI models · Mistral AI

Mistral Releases First Code Large Model Codestral ↗

Mistral has released its first code large model, Codestral, featuring 22B parameters, a 32k context window, and support for over 80 programming languages. It offers two usage options: open-source weights (under the Mistral AI Non-Production License) and commercial licensing.

AI models · Mistral AI

Mistral AI Releases Non-Production License Agreement MNPL ↗

Mistral AI has released the Mistral Non-Production License (MNPL), which allows developers to use the technology for non-commercial purposes and research; the Codestral model is being released under this new license effective immediately; the company also stated that it will continue to release some models and code under Apache 2.0, forming a dual-license approach.

Practical tips · Hugging Face Blog

Hugging Face Releases TGI Benchmarking Tool ↗

Hugging Face's official blog introduced its benchmarking tool for Text Generation Inference (TGI), which is used to quantitatively analyze the impact of different configurations (especially batch size) on prefill latency (TTFT).

AI models · Hugging Face Blog

TII Releases Falcon 2-11B Large Language Model and Multimodal Vision-Language Model ↗

The Technology Innovation Institute (TII) of the UAE released Falcon 2-11B, the first model in the Falcon 2 series. It includes an 11B-parameter pure-text large language model and a vision-language model (VLM) built upon it, supporting English and ten languages including Spanish and French. The training data exceeds 5000B tokens, with a context window of 8192, under an Apache 2.0-derived license.

Research · Hugging Face Blog

Meta 发布 CyberSecEval 2：面向大语言模型的网络安全风险综合评测框架 ↗

Meta 发布 CyberSecEval 2，一个开源的 LLM 网络安全风险评测框架，包含五类测试：不安全编码实践生成、提示注入攻击易感性、协助网络攻击请求的合规性、代码解释器滥用倾向、自动化进攻性网络安全能力。所有代码开源，支持社区运行并提交结果至排行榜。

Source-language fallback
AI products · Hugging Face Blog

Hugging Face Spaces Launches Dev Mode ↗

Hugging Face Spaces has introduced the Dev Mode feature, which supports one-click connection to VS Code or SSH for directly editing Space code without requiring git pushes or container rebuilds; this feature is currently in beta and available only to PRO subscribers.

AI products · Hugging Face Blog

Hugging Face and Dell Jointly Launch Dell Enterprise Hub ↗

Hugging Face and Dell jointly launched the Dell Enterprise Hub, offering a one-stop experience for locally training and deploying open-source large language models (such as Llama 3, Mixtral, and Gemma) on Dell hardware.

AI models · DeepSeek 官方更新

DeepSeek-V2-0517 Released: Upgraded Instruction Following and JSON Output Capabilities ↗

The deepseek-chat model has been upgraded to DeepSeek-V2-0517. The IFEval Prompt-Level accuracy increased from 63.9% to 77.6%. Instruction-following capabilities in the system prompt area via API have been optimized, enhancing immersive translation and RAG experiences. The JSON parsing rate increased from 78% to 85%, reaching 97% after introducing regular expressions.