Archive
新模型、開放權重、多模態、語音、影像與模型架構。
GitHub Repo
IFM has released the K2 Horizon model family, ranging from 0.9B to 375B parameters. The 36B version extends MoE-style routing to the attention mechanism’s value projections. Weights and deployment recipes are now available, but the organization’s claim of being “fully open source” still runs ahead of what it has actually delivered.
AI 安全
Microsoft recommends checking both hardware-rooted remote attestation and AI artifact provenance before releasing model weights, data, or credentials to edge devices. Models should only propose actions; actual authorization should be handled by a predictable external policy layer to prevent prompt injection from directly gaining system privileges.
最新模型
The new method uses a large teacher model to synthesize task data, then trains a function-specific LoRA for each task so subsequent requests do not need to call a remote model. It achieves 83.6% semantic accuracy on a difficult benchmark, but compilation still requires a teacher API, about 40 GB of accelerator memory, and output validation.
代理工具
The Copilot CLI research preview no longer selects just one model for each task. Instead, it routes tasks among three workflows: direct execution, staged escalation, and cross-model review. GitHub’s offline tests indicate that some workloads may reduce estimated costs, but the router, quality gates, and internal benchmark have not been disclosed.
代理工具
Researchers from Tencent Hunyuan and collaborators make terminal environments progressively harder across generations based on scenario novelty, skill rarity, and execution length, without continually using rollouts from the model being trained to identify weaknesses. In a 200-step reinforcement learning experiment, the method outperformed co-evolution baselines, but the findings currently come only from the authors, and a complete reproduction package has not yet been released.
最新模型
Google has moved Gemini 3.8 Flash into production, offering a 1 million-token context window, three thinking levels, and built-in tools, while making it the default model for Antigravity managed agents. Coding benchmark results have improved, but most figures still come from Google, and the model card also reveals a relative decline in non-English safety evaluations.
AI 安全
Cloudflare’s early-preview service combines source code analysis results with actual routes, traffic, WAF status, and attack events, then uses GPT-5.6 Cyber to validate vulnerabilities and proposed fixes. The model does not directly modify the environment, but data is sent to OpenAI for inference, and Cloudflare has not disclosed false-positive or remediation success rates.
電腦視覺
Scal3R retains a frozen CUT3R or STream3R backbone, using lightweight query tokens to estimate relative poses against multiple historical keyframes before passing them to online pose graph optimization. The authors report a mean ATE on KITTI more than 60% lower than the strongest online competitor, but the results use Sim(3) alignment and therefore should not be interpreted as preserving true scale.
最新模型
Microsoft Research’s streaming speech model interleaves audio chunks, brief lookahead, and the existing transcript to directly output “who said what.” The public release supports ten languages and custom hotwords, but expected speaker-attribution latency remains two seconds, and performance degrades during prolonged overlapping speech.
推論系統
KAIST and Google DeepMind have introduced Declarative Attention, which allows existing models to switch among global, focused, and local attention modes within their reasoning text. The authors reduced attended tokens by 31.1% to 52.0% across two models, though the speed gains remain model-based estimates and accuracy declined slightly.
最新模型
Zhipu AI has unveiled GLM-5.3-Flash, a native multimodal model with 320B total parameters and 18B active parameters per token, releasing its weights under the MIT License. The new architecture is claimed to require three times less attention computation and 4.4 times less KV cache than GLM-5.3, but “Flash” does not mean it can be easily deployed on a typical workstation.
推論系統
Uno attaches lightweight diffusion weights to an autoregressive LLM, then uses Ψ-Spec to propose multiple tokens at once while preserving the original model’s sampling distribution. The authors report speedups of up to 3×, though the current figures are limited to the research team’s specific models, hardware, and inference engine.