Archive
新模型、開放權重、多模態、語音、影像與模型架構。
代理工具
P-Bench requires agents to select statistical methods, execute R code, and report p-values themselves, rather than merely checking whether the code runs. The 14B model leads the listed open models on the strict metric for difficult tasks, but the public repository currently contains only documentation.
推論系統
A community deployment recipe running on a single DGX Spark uses CUDA Graph and a one-layer MTP speculative decoder to increase Ling-3.0-flash INT4 generation speed from 20.8 to 38.7 tokens/s. More importantly, it finds that mainline vLLM does not yet support the V3 architecture: forcing the model through an older attention path may not raise an error but can produce plausible-looking yet incorrect content.
最新模型
MiniMax H3 packages text, images, video, and audio into a single sequence and jointly predicts visual and audio latents, producing up to 768p video at 24 FPS with 32 kHz stereo audio. The official workflow claims regeneration at up to 2K, but prompt orchestration and the high-resolution module remain available only through the hosted API.
代理工具
Scale AI treats prompts, tools, memory, and control flow as model-modifiable code, then measures real-world gains using a hidden test set. Across 111 experiments, native coding agents did not consistently outperform a shared harness, while the best validation scores generally proved overly optimistic.
代理工具
Rather than retaining every executable synthetic task, CalibForge repeatedly revises tasks until different solver agents exhibit predefined success and failure patterns. Two models fine-tuned on calibrated trajectories improved across three coding-agent benchmarks, but full training required 64 NVIDIA H20 GPUs, and some results were based on only a single run.
AI 安全
A new benchmark uses four paired contexts to distinguish whether models correctly evaluate evidence or simply ignore all external information. By changing only the configuration of DPO preference data, SCOPE reduces answer flips caused by misleading contexts while preserving accuracy in correct and irrelevant contexts.
最新模型
An expanded version of Low Frequency Trap uses 2,190 programmable videos to measure event count and frequency separately. Gemini 3.6 Flash’s ball-bounce counting accuracy improves with increased sampling, but its timestamp-sequence F1 score is only 3.7%.
最新模型
A new study uses counterfactual cropping to test six open models, distinguishing between “calling a tool” and “the tool’s returned content actually changing the answer.” Most positive gains are concentrated in a small number of well-calibrated trajectories, suggesting that final accuracy alone may overstate visual tool capabilities.
代理工具
During training, EnvACE does not query external environments. Instead, the same policy alternates between generating tool actions and imagined environment responses. This reduces reliance on executable training environments, but also directly absorbs simulation errors into the agent’s parameters.
最新模型
Rather than treating teacher–student differences at every token equally, DASH adjusts the scope of supervision based on how divergence changes across an entire reasoning sequence. All three Qwen3 sizes outperform OPSD under the same conditions, but selecting the best checkpoint by average test-benchmark score may overestimate generalization gains.
推論系統
PaDoc decodes the layout sequence and each region’s content in parallel while sharing the page prefix, avoiding serialization of the entire document into a single path. Its 2.1B model maintains a 94.24 Overall score on OmniDocBench, though the performance figures come from a single GPU under specific scheduling settings.
硬體
The open-source esp32-ai project uses 4-bit quantization and tiered memory allocation to keep roughly 25 million embedding parameters in flash, reading only the rows needed for each token. The model can generate text offline on a microcontroller costing under $10, though its capabilities remain limited to short stories and single-domain question answering.