Latest

Training on internalized skill workflows lets a 35B open-weights model beat far larger proprietary models on agentic benchmarks, shifting the edge from raw model scale to curated environment data. Builders running Claude Code or Codex harnesses can reuse the released environments and trajectories for their own SFT or RL pipelines. The results give Qwen-based deployments a concrete benchmark point against Claude Sonnet 4.6 and DeepSeek V4 Pro.

The key finding is that skills can be internalized as model weights rather than injected at inference time, since the fine-tuned agent beats skill-assisted base models even without the skills provided. Agent engineers and fine-tuning practitioners can reuse the released environments and trajectories for SFT or RL with outcome-based rewards. The result also shows a 35B open Qwen-based model outscoring much larger closed models on SkillsBench, a useful cost comparison for teams building coding and tool-use agents.

The core result is that a quantized local model can match a cloud API on Vietnamese compliance extraction, removing the data-sovereignty objection that blocks LLM use under privacy law. Compliance teams and legal-tech builders in Vietnam face a 1 January 2026 deadline, and this shows a single 24 GB GPU suffices for the workload. The unmeasured value-level precision and mid-range token coverage are caveats to weigh before deployment.

Builders running memory-augmented agents on open Qwen3 models can adopt this approach, since it beat strong memory baselines on LoCoMo and PersonaMem without adding retrieval routing or iterative search.

This research introduces a new AI model for evaluating speech synthesis, which could be useful for developers working on TTS systems.

This provides a new benchmark for evaluating how well large language models can handle complex, multi-step scientific reasoning in interdisciplinary fields like biophysics, which is useful for researchers comparing model performance on specialized tasks.

This research provides a concrete improvement for using AI in financial document processing, showing how to make VLMs more reliable for key-value extraction.

This research provides a new method for aligning large language models with human preferences, which is a key challenge in AI development.

This research presents a new approach to action tokenization that could improve the performance of AI models in robotics and other fields by better preserving the physical relationships between actions.

This research introduces a more efficient way to route skills in large language models, potentially reducing the need for large external parameter counts.

For those working with camera traps in the field, this study provides crucial data on how well current models can identify species in real-world conditions, showing significant performance drops on low-quality images.

This is a valuable resource for researchers and developers working on multimodal AI, as it provides a new benchmark to evaluate topological reasoning capabilities in large language models.

This paper presents a new parameterization for per-token gating in on-policy knowledge distillation, which is a key technique for training smaller models to mimic larger ones. It provides a framework for comparing different gating designs.

TRACE demonstrates that simulation-based verification can make diagnostic reasoning tasks scalable for reinforcement learning, outperforming larger models with fewer tool calls.

Maverick enables private LLM inference with 45x speedup over local processing, addressing privacy concerns for users who must use third-party providers.

The study reveals optimal data distribution for multi-domain models, helping developers balance coverage during mid-training to maximize final performance.

The study provides concrete metrics for evaluating AI-generated industrial data, crucial for manufacturers adopting LLMs in Industry 4.0 implementations.

This quantization method enables smaller, faster hybrid LLMs without significant performance loss, making large models more accessible for resource-constrained environments.

Developers can implement SafeEvolve to improve agent safety without sacrificing utility, achieving better safety-utility tradeoffs than existing methods.

MT-SDPO improves multi-domain LLM performance by 14.79 points on Qwen3-8B, offering better domain balance than single-teacher approaches.

Global quantization granularity beats selective layer repair by up to 52 accuracy points, offering a better default for LLM deployment.

The study provides concrete guidance for selecting LLMs in ontology engineering, showing model architecture matters more than size for specific tasks.

BLOOM-WILT improves automated LLM auditing efficiency, making it valuable for companies deploying AI systems in ASEAN markets.

The method improves reasoning accuracy in existing models without retraining, making it valuable for developers working with constrained resources in ASEAN.