Tag: research
All the articles with the tag "research".
-
MIT's DisCIPL Enables Small LMs to Match GPT-4o in Reasoning Efficiency
• 1 min readMIT CSAIL's DisCIPL uses LLMs for planning, delegating to small models for complex reasoning—outperforming GPT-4o with less compute.
Read more -
Anthropic Releases Claude Opus 4.5, Outperforms Humans in Engineering Tests
• 1 min readAnthropic's Claude Opus 4.5 beats all human candidates in internal engineering tests, escalating the AI capability race with Gemini 3 and ot
Read more -
OpenAI Launches GPT-5.2 After 'Code Red' Sprint
• 1 min readOpenAI releases GPT-5.2 with enhanced reasoning, coding, and long prompts after internal 'code red' to counter Google's Gemini 3. Now in Cha
Read more -
Microsoft’s Agent Lightning: Add Reinforcement Learning to Agents Without Rewriting Code
• 1 min readAgent Lightning lets existing LLM-based agents learn via RL with minimal code changes, improving task accuracy across real workloads.
Read more -
OpenAI's Enhanced Pretraining Strategy 'Garlic' Signals Next GPT Upgrade
• 1 min readOpenAI's 'Garlic' project advances GPT-5.2/5.5 with improved pretraining and bug fixes, aiming for early 2026 release.
Read more -
Cognizant AI Lab Achieves Breakthrough in LLM Fine-Tuning with Evolution Strategies
• 1 min readCognizant pioneers evolution strategies to fine-tune billion-parameter LLMs, surpassing reinforcement learning.
Read more -
DriveMLM: Multi-Modal LLM Framework Enhances Autonomous Driving with Human-Like Reasoning
• 1 min readDriveMLM integrates multi-modal inputs to improve autonomous vehicle planning and explainability.
Read more -
AI Encourages Mainstream Adoption of Formal Verification via LLM Assistance
• 1 min readLLMs improve formal verification, making it easier and poised for widespread use in software correctness.
Read more -
Open-Weight Mistral Large 3 Model Advances Multimodal and Efficient LLM Inference
• 1 min readMistral Large 3 is a powerful open-weight LLM with 41B active params, supporting text, images, and low-precision inference.
Read more -
AI Techniques to Address Hardware-Software Challenges in Quantum Computing
• 1 min readResearchers apply AI methods to tackle hardware-software integration issues in practical quantum computing development.
Read more -
Emergence of Post-LLM Architectures for Next-Gen AI
• 1 min readPost-LLM architectures combine LLMs with new techniques to surpass current limitations and enable versatile AI.
Read more -
Andreessen Horowitz and OpenRouter Reveal Usage Patterns from 100 Trillion Tokens
• 1 min readAnalysis of 100T tokens exposes growing complexity and agentic multi-model AI use in real-world LLM sessions.
Read more
