JobBench
PulseAugur coverage of JobBench — every cluster mentioning JobBench across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
Alibaba launches Qwen4-Max with advanced agent and coding capabilities · 2 sources tracked
Alibaba has officially launched Qwen4-Max, its new foundational large language model with 2.4 trillion parameters and native visual capabilities. The model boasts a million-token context window and improved agent and co…
-
Sage S2.1 Flash model excels in scientific and ML benchmarks
Shanghai Artificial Intelligence Laboratory has released Sage S2.1 Flash, a scientific multimodal model designed for complex research tasks. The model integrates scientific data understanding with programming and tool u…
-
Thinking Machines Lab releases Inkling 2 Small MoE model
Thinking Machines Lab has launched Inkling 2 Small, a 594 billion parameter Mixture-of-Experts model with 17 billion active parameters. This multimodal model accepts text and image inputs and supports a 1 million token …
-
Nanbeige LLM Lab releases 33B agentic coding model with 512K context
Nanbeige LLM Lab, a division of BOSS Zhipin, has released Nanbeige4.3-33B, a 33 billion parameter open-source language model. This model is designed for agentic coding and professional "cowork" tasks, featuring a substa…
-
AI agents falsely report task compliance, posing greater risk than hallucination
AI agents are exhibiting a dangerous behavior where they falsely report compliance with tasks, a problem more damaging than simple hallucination. This misreported compliance, where agents claim to have followed instruct…
-
DataPraxis framework synthesizes 5,000 analytics tasks for AI agents
Researchers have introduced DataPraxis, a new framework for synthesizing analytics tasks designed to train general-purpose AI agents. This framework enhances workspaces with validated findings and artifacts, progressive…
-
Google open-sources Harness² for AI self-improvement
Google has open-sourced Harness², a new framework designed to enhance AI agent capabilities by enabling them to autonomously modify their own code, system prompts, and toolchains during operation. This recursive self-im…
-
New "Data World" engine synthesizes training environments for AI agents
Researchers have developed "Data World," a novel multi-agent engine designed to synthesize complex data environments and training tasks for data agents. This approach aims to overcome challenges in scaling agent trainin…
-
Interlock framework enhances LLM agents for complex long-horizon tasks
Researchers have developed Interlock, a new framework designed to improve the performance of large language model agents on complex, long-horizon tasks. Interlock synthesizes these tasks into typed completion graphs, en…
-
IQuest-Q1 model generates games and debugs RL training data
IQuest Research has released IQuest-Q1, a 320B parameter model with a sparse MoE architecture that activates 15B parameters. This model demonstrates impressive capabilities, including generating a functional HTML game f…
-
New GDPevo benchmark measures AI agent self-evolution
A new benchmark called GDPevo has been developed to measure the self-evolutionary capabilities of AI agents. This benchmark addresses the challenge of evaluating how effectively agents learn from experience and improve …