TradingAgents: Multi-Agents LLM Financial Trading Framework

A multi-agent framework using large language models for stock trading simulates real-world trading firms, improving performance metrics like cumulative returns and Sharpe ratio.

https://arxiv.org/abs/2412.20138

https://hjfy.top/arxiv/2412.20138

https://github.com/tauricresearch/tradingagents (120972)

VibeVoice Technical Report

VibeVoice synthesizes long-form multi-speaker speech using next-token diffusion and a highly efficient continuous speech tokenizer, achieving superior performance and fidelity.

https://arxiv.org/abs/2508.19205

https://hjfy.top/arxiv/2508.19205

https://github.com/microsoft/VibeVoice (56385)

TokenEconomizer: Experience-Driven Black-Box Cost Control for Coding Agents

TokenEconomizer reduces recurring token costs of coding agents across sequential real-world tasks by pruning irrelevant tools, injecting micro-skills, and interrupting failure loops.

https://arxiv.org/abs/2610.22732

https://hjfy.top/arxiv/2610.22732

https://github.com/microsoft/TokenEconomizer (4866)

VisSearch-Bench: Evaluating Long Horizon Visual Search in Multimodal Agents

VisSearch-Bench evaluates multimodal agents on visually intensive search tasks that require integrating visual context with external information, showing that even the best system fully solves only 26.47% of tasks.

https://arxiv.org/abs/2611.04894

https://hjfy.top/arxiv/2611.04894

https://github.com/antins-labs/VisSearch-Bench-v1 (57)

WeVisOne-VL: Unified Visual Perception from Objects to Pixels

WeVisOne-VL unifies object-level and pixel-level visual perception via structured token prediction, supporting detection, localization, counting, OCR, segmentation, and matting with a lightweight pixel decoder and multi-stage training.

https://arxiv.org/abs/2610.31329

https://hjfy.top/arxiv/2610.31329

https://github.com/Tencent/WeVisOne-VL (1048)

The Other Half of the Memory Wall: Serving 35B MoEs from SSD with Trained Routing Prediction

Edge0 enables efficient mixture-of-experts inference on consumer hardware by predicting expert routing ahead of time and using recovery adapters to maintain quality with minimal active memory.

https://arxiv.org/abs/2609.18063

https://hjfy.top/arxiv/2609.18063

https://github.com/Edge0-AI/edge0 (9111)

SpecPilot: Autonomous Algorithm Discovery and Systems Optimization for Speculative Decoding

SpecPilot autonomously discovers deployment-specific algorithms and system optimizations to maximize speculative decoding speedups across diverse models, drafters, and serving regimes without quality loss.

https://arxiv.org/abs/2611.02501

https://hjfy.top/arxiv/2611.02501

https://github.com/StargazerX0/SpecPilot (168)

ARIS: Autonomous Research via Adversarial Multi-Agent Collaboration

ARIS is an open-source research harness that uses cross-model adversarial collaboration to ensure reliable long-term research outcomes through coordinated execution, orchestration, and assurance layers.

https://arxiv.org/abs/2605.03042

https://hjfy.top/arxiv/2605.03042

https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep (18775)

TailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing

TailSteer, a final-stage training framework, reduces cumulative visual drift in multi-turn editing by keeping early denoising dynamics close to the base model and learning tail-focused preservation corrections from on-policy rollouts, preserving pretrained generative and editing abilities.

https://arxiv.org/abs/2611.06741

https://hjfy.top/arxiv/2611.06741

https://github.com/Lyricccco/tailsteer (37)