JudgeMyAICHIMERA Dataset Delivers 4B Model Reasoning Performance on Par with 235B Models
CHIMERA synthetic dataset enables 4B models to achieve reasoning performance comparable to 235B models.
GEPA and SkillOpt Optimization: When Reflective Prompting and Skill Optimization Help on Locally Served Reasoning Models
GEPA and SkillOpt optimization effects on locally served reasoning models. Task-dependent gains and validation challenges.
GroundingPI Sets New Benchmark for Physical Intelligence with 73.68% Accuracy Across 34 Grounding Tasks
GroundingPI achieves 73.68% accuracy across 34 grounding benchmarks, outperforming GPT-6 Astra by 2.14%.
Sphinx: The Git Hook for Agent-Assisted Development
Sphinx is a Git hook that quizzes developers on their AI-generated code before shipping.
HTN-Orchestrate: How Hierarchical Task Networks Boost Multi-Agent Workflow Accuracy by 2.45–4.69 Percentage Points
HTN-Orchestrate outperforms ReAct and Agent Team by 2.45–4.69% on BrowseComp and SUPERChem benchmarks.
M$^3$R-Bench Reveals Critical Evidence-Mapping Gaps in Multimodal Metaphor Understanding
M$^3$R-Bench exposes critical evidence-mapping failures in multimodal metaphor understanding. Learn how M$^3$R-Reasoner addresses these gaps.
Gnani-Evon-v3.3-30B-A3B: A 5.2-Point Indic Reasoning Gain at 1.7-Point English Cost
Analysis of Gnani-Evon-v3.3-30B-A3B’s 5.2-point Indic reasoning gain and 1.7-point English cost.
Evaluating Grok 4.7’s Long-Horizon Agent Capabilities on Amazon Bedrock
Evaluating Grok 4.7’s long-horizon agent capabilities on Amazon Bedrock.
Global Cross-Region Inference: Unlocking OpenAI GPT-6 Models in Asia Pacific
Explore global cross-region inference for OpenAI GPT-6 models in Asia Pacific with Amazon Bedrock.
ZenGen Framework: Social Intelligence Benchmarks Reveal 28% Accuracy Headroom for LLMs
ZenGen framework reveals 28% accuracy gap in social intelligence for LLMs. Learn about SoMBench and Actio deployment architecture.