模型
ChatGPT Lockdown Mode Introduced
OpenAI launches Lockdown Mode and Elevated Risk labels in ChatGPT. These tools protect organizations from prompt injection and AI data exfiltration attacks.
OpenAI News · 216 天前
February 13, 2026
模型
ChatGPT Lockdown Mode Launched
OpenAI introduced Lockdown Mode and Elevated Risk labels in ChatGPT. These features help organizations defend against prompt injection attacks.
OpenAI News · 216 天前
模型
ChatGPT Launches Lockdown Mode and Risk Labels
OpenAI introduces Lockdown Mode and Elevated Risk labels in ChatGPT. These features help organizations defend against prompt injection attacks.
OpenAI News · 216 天前
基建
Scaling Access Beyond Rate Limits for Codex and Sora
OpenAI has built a real-time access system for Codex and Sora that surpasses traditional rate limits. The system combines rate limits, usage tracking, and credits to enable continuous access.
OpenAI News · 216 天前
基建
Scales Codex Sora Access
OpenAI built a real-time access system for Codex and Sora. It integrates rate limits, usage tracking, and credits.
OpenAI News · 216 天前
基建
Scaled Access for Codex and Sora
OpenAI built a real-time access system beyond traditional rate limits. It integrates usage tracking and credits for continuous access to Codex and Sora.
OpenAI News · 216 天前
基建
Scaled Access for Codex Sora
OpenAI built a real-time access system for Codex and Sora. It integrates rate limits, usage tracking, and credits.
OpenAI News · 216 天前
開源
OpenAI Launches GABRIEL Toolkit
OpenAI has released GABRIEL, a new open-source toolkit. It uses GPT to convert qualitative text and images into quantitative data.
OpenAI News · 216 天前
開源
GABRIEL Toolkit for Social Science
OpenAI launched GABRIEL, an open-source toolkit. It uses GPT to convert qualitative text and images into quantitative data.
OpenAI News · 216 天前
開源
GABRIEL Social Science Toolkit
OpenAI launches GABRIEL, an open-source toolkit powered by GPT. It converts qualitative text and images to quantitative data.
OpenAI News · 216 天前
開源
GABRIEL Scales Social Science
OpenAI releases GABRIEL, an open-source toolkit powered by GPT. It converts qualitative text and images into quantitative data for large-scale analysis.
OpenAI News · 216 天前
商業
WEF Urges Action on AI Uncertainties
World Economic Forum in Davos hosted over 200 AI sessions for leaders. Discussions covered AI's role in geopolitics, productivity, job disruption, ethics, governance, and environmental impact.
SCMP Technology · 216 天前
商業
AI Leads Davos with 200 Sessions
Davos World Economic Forum hosted over 200 AI sessions for leaders. Key themes include geopolitical rivalry, productivity boost, job disruption, ethics, and environmental impact.
SCMP Technology · 216 天前
商業
SoftBank Borrows $27B for OpenAI Bet
SoftBank Group added $27 billion in new debt during Q4 to fund its heavy investment in OpenAI. The company also sold Nvidia shares for $5.8 billion and T-Mobile shares for over $3.5 billion.
cnBeta (Full RSS) · 216 天前
基建
China's Sustainable AI Data Centers in Sea and Space
China is innovating AI data center sustainability by deploying them underwater, in mountains, and in space to cut energy and water use. Unlike Microsoft's failed underwater experiment five years ago, China has implemented all three methods.
SCMP Technology · 216 天前
基建
China Pioneers Underwater, Mountain, Space Data Centers
China deploys AI data centers underwater, in mountains, and in space for sustainability. These reduce energy and water use amid AI boom.
SCMP Technology · 216 天前
開源
OpenClaw 2026.2.12: 40 Fixes, New Models, Cron Overhaul
OpenClaw 2026.2.12 delivers the largest security release ever with 40 fixes. It adds support for GLM-5 and MiniMax M2.5 models plus IRC integration.
OpenClaw.report · 216 天前
開源
OpenClaw 2026.2.12: 40 Fixes and More
OpenClaw 2026.2.12 delivers 40 security fixes, its largest ever. Adds GLM-5, MiniMax M2.5 model support, IRC integration, and stabilized cron scheduler.
OpenClaw.report · 216 天前
開源
OpenClaw 2026.2.12 Packs 40 Fixes
OpenClaw version 2026.2.12 includes 40 security fixes, the largest ever. It adds GLM-5 and MiniMax M2.5 model support, IRC integration, and a stabilized cron scheduler overhaul.
OpenClaw.report · 216 天前
編程
OpenAI Launches Speedy Codex-Spark Model
OpenAI released GPT-5.3-Codex-Spark, a lightweight version of its Codex intelligent agent programming tool. This slimmed-down model prioritizes extreme inference speed for rapid iteration scenarios.
cnBeta (Full RSS) · 216 天前
編程
OpenAI Launches Fast Codex-Spark Variant
OpenAI released GPT-5.3-Codex-Spark, a lightweight version of its Codex programming tool. It targets rapid iteration with extreme inference speed.
cnBeta (Full RSS) · 216 天前
編程
OpenAI Launches Speedy Codex-Spark Variant
OpenAI released GPT-5.3-Codex-Spark, a lightweight version of its Codex intelligent programming tool. Designed for extreme inference speed, it targets rapid iteration scenarios.
cnBeta (Full RSS) · 216 天前
音訊
Voxtral Realtime Streaming ASR
Voxtral Realtime achieves Whisper-quality transcription at 480ms latency via end-to-end streaming training. Features causal audio encoder and Ada RMS-Norm.
ArXiv AI · 216 天前
圖像
V2G Transforms CAD into Auditable Graphs
V2G pipeline converts CAD diagrams into property graphs, capturing component topology and connectivity overlooked by pixel-based MLLMs. It delivers major accuracy gains on electrical schematic compliance benchmarks where top MLLMs fail.
ArXiv AI · 216 天前
模型
TSR Boosts Multi-Turn RL for LLM Agents
TSR introduces trajectory-search rollouts to enhance multi-turn reinforcement learning for LLM agents. It uses lightweight tree-style search for high-quality trajectories, improving rollout generation and stabilizing training.
ArXiv AI · 216 天前
模型
TRACER Aggregates Risks in Agent Trajectories
TRACER is a trajectory-level uncertainty metric for tool-using agents, combining surprisal, repetition, and coherence signals with tail-focused aggregation. Improves AUROC by 37% and AUARC by 55% on tau^2-bench for failure prediction.
ArXiv AI · 216 天前
模型
ThinkRouter Boosts Reasoning Efficiency
ThinkRouter introduces confidence-aware routing between latent and discrete spaces for efficient AI reasoning. It switches to discrete tokens during low-confidence steps to reduce noise from latent embeddings.
ArXiv AI · 216 天前
編程
Text2GQL-Bench: New Graph Query Benchmark
Text2GQL-Bench introduces a unified benchmark for Text-to-Graph-Query-Language systems with 178,184 question-query pairs across 13 domains and multiple GQLs. It features a scalable dataset generation framework and a multi-metric evaluation including grammatical validity, similarity, semantic alignment, and execution accuracy.
ArXiv AI · 216 天前
模型
Surveying Multi-Agent Communication Paradigms
This survey frames multi-agent communication via the Five Ws, tracing evolution from MARL's hand-designed protocols to emergent language and LLM-based systems. It highlights trade-offs in interpretability, scalability, and generalization across paradigms.
ArXiv AI · 216 天前
應用
SemaPop: Semantic Population Synthesis
SemaPop uses LLMs for semantic-conditioned population synthesis, deriving personas from surveys. Integrates with WGAN-GP for statistical alignment and behavioral realism.
ArXiv AI · 216 天前