FrontierSWE
PulseAugur coverage of FrontierSWE — every cluster mentioning FrontierSWE across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
MaxxBench reveals coding AI struggles with requirement changes
A new benchmark called MaxxBench has been developed to evaluate the adaptability and generalization capabilities of frontier coding agents. MaxxBench introduces small mutations to existing tasks, such as changing file n…
-
Google DeepMind launches Gemini 4 Carbon for cybersecurity and coding
Google DeepMind has unveiled Gemini 4 Carbon, a new frontier model designed for advanced cybersecurity defense and complex coding tasks. This model reportedly outperforms its predecessor, Gemini 4 Argon, and other leadi…
-
Spine runtime improves AI coding agents by structuring subtask context
Researchers have developed Spine, a new runtime and interface designed to manage long-horizon coding tasks by structuring subtask context. Unlike sequential history recording or recursive subagent delegation, Spine orga…
-
Stateless AI agents outperform those with large context windows, study finds
A new paper from Stanford and other universities challenges the assumption that larger context windows and longer agent runs inherently lead to better performance. The research found that for automated research systems,…
-
inclusionAI releases Ling-3.1-flash with 560B parameters and 1M context window
inclusionAI has released Ling-3.1-flash, an advanced hybrid reasoning model with 560 billion total parameters and 25 billion active parameters per token. This new model builds upon its predecessor, Ling-3.0-flash, by en…
-
Moonshot launches Kimi K3 with 2.8T parameters and 1M context window
Moonshot has launched its Kimi K3 model, a 2.8-trillion-parameter Mixture-of-Experts model with a context window of over 1 million tokens. The model features a new Kimi Delta Attention mechanism, which combines linear a…
-
Anthropic's Claude Fable 5.1 cuts cache costs but raises per-task prices
Anthropic's Claude Fable 5.1 has been released, featuring a significant reduction in cache read costs by 75%, making it cheaper for tasks involving repeated context. However, the cost per task has increased by approxima…
-
New framework enables LLM agents to autonomously develop and improve software
Researchers have introduced Harness-of-Harness (HoH), a framework designed to enhance autonomous software development by enabling LLM-based coding agents to continually improve software. HoH structures development into …
-
Bespoke Labs seeks researcher for long-horizon agent benchmarks
Bespoke Labs is seeking a researcher to develop and evaluate reinforcement learning environments and benchmarks for long-horizon agent tasks. The role requires demonstrated experience with multi-step reasoning agents, s…
-
China's Z.ai releases low-cost GLM 5.2, challenging US AI model dominance
Z.ai, a Beijing-based lab, has released GLM 5.2, an open-weights AI model that significantly undercuts the cost of leading U.S. frontier models. While GLM 5.2's API costs are a fraction of those for models like Anthropi…
-
Moonshot AI releases Kimi K3 with 1M context, open weights, and competitive pricing
Moonshot AI has released Kimi K3, an open-weights model with approximately 2.8 trillion parameters and a 1 million token context window. The model achieved strong performance on benchmarks like Terminal-Bench 2.0, Front…
-
Zhipu AI releases GLM-5.2 with 1M context window, challenging top proprietary models
Zhipu AI has released GLM-5.2, a 744B-parameter Mixture-of-Experts model featuring a 1 million token context window and MIT-licensed weights. This model achieves a high ranking on the BenchLM leaderboard and demonstrate…
-
China's GLM-5.2 challenges GPT-5.5 and Claude Opus on coding benchmarks
Zhipu AI's GLM-5.2, a Chinese frontier model, has reportedly achieved strong performance on coding benchmarks, surpassing OpenAI's GPT-5.5 and Anthropic's Claude Opus 4.7. On the FrontierSWE benchmark, GLM-5.2 scored 74…
-
Z.ai releases GLM-5.2, setting new open-source benchmark for long-context AI
Z.ai has released GLM-5.2, an open-source language model with a 1 million token context window, positioning it as a strong contender in long-horizon tasks and coding benchmarks. The model features an improved architectu…