SWE-Marathon
PulseAugur coverage of SWE-Marathon — every cluster mentioning SWE-Marathon across labs, papers, and developer communities, ranked by signal.
- developed by Muse Wonder 95%
- instance of Meta Model API 90%
- instance of GLM-5.2 90%
- competes with GPT -6 Astra 70%
- competes with Claude (Opus 5.5 70%
- instance of Muse Wonder 70%
- used by Muse Wonder (Full) 70%
- competes with Meta Model API 70%
- used by Meta Model API 70%
- used by Muse Code 70%
- competes with GLM-5.2 70%
- competes with Muse Wonder 60%
3 day(s) with sentiment data
-
Moonshot's Kimi K3.1 Pro shows major gains with enhanced post-training
Moonshot has released Kimi K3.1 Pro, an updated version of its Kimi K3 model, which shows significant improvements in coding and hallucination benchmarks due to enhanced post-training. While the base model architecture …
-
Moonshot AI releases Kimi K3.1 Pro with 1M context window and 2.8T parameters
Moonshot AI has launched Kimi K3.1 Pro, a new model in its K3.1 series, featuring 2.8 trillion total parameters with 104 billion active parameters in a Mixture-of-Experts architecture. This model boasts a 1-million-toke…
-
Meta's Muse Wonder agent faces scrutiny over 3D task failure and benchmark claims
Meta's new Muse Wonder coding agent has sparked debate in the developer community following a 3D programming test that revealed significant limitations. While Meta's official benchmarks showed strong performance, a comm…
-
Meta's Muse Wonder AI models launch amid mixed reviews and trust concerns
Meta Superintelligence Labs has released its first frontier-class AI models, Muse Wonder and Muse Wonder Full, achieving competitive scores on various benchmarks, including the Artificial Analysis Intelligence Index whe…
-
Meta unveils Muse Wonder AI models for agents
Meta has announced its new AI models, Muse Wonder and Muse Wonder Full, which are described as their most capable to date and successors to the Muse Spark series. These models are designed for personal and work AI agent…
-
Meta launches Muse Wonder models with Active Focus and Memory
Meta has introduced Muse Wonder and Muse Wonder Full, its latest models designed for personal and professional agentic tasks. These models feature 'Active Memory' for context retention and 'Active Focus,' which enables …
-
Meta's Muse Wonder AI challenges Anthropic's Opus 5.5 in coding benchmarks
Meta has launched Muse Wonder, a new AI model that competes with Anthropic's Claude Opus 5.5, particularly in long coding tasks. While Meta's internal benchmarks show Muse Wonder outperforming Opus 5.5 in certain areas …
-
Meta's Muse Wonder AI launches at Opus 5.5 prices, trails in independent tests · 6 sources tracked
Meta has launched its new frontier AI models, Muse Wonder and Muse Wonder Full, designed for agentic tasks. While Meta's internal benchmarks show Muse Wonder Full outperforming competitors like Claude Opus 5.5 and GPT-6…
-
DeepSWE Meta Muse tops SWE-Marathon benchmark, outperforming Claude Opus 5.5 and GPT-6 Astra
A new model named DeepSWE Meta Muse has demonstrated superior performance on the SWE-Marathon benchmark, achieving a score of 57.5%. This score surpasses that of Claude Opus 5.5, which scored 46.3%, and GPT-6 Astra, whi…
-
Meta returns to AI frontier with Muse Wonder models
Meta has launched its new frontier-class AI models, Muse Wonder and Muse Wonder Full, marking a significant return to the forefront of AI development. These models demonstrate advanced capabilities in coding, long-conte…
-
Zhipu AI's GLM-5.3 shows China's AI advancement, bypassing US restrictions
Zhipu AI has released its new GLM-5.3 model, which significantly improved performance on the SWE-Marathon benchmark. This advancement demonstrates China's capability in developing advanced AI models, potentially circumv…
-
Zhipu AI releases GLM-5.3, excelling in coding and security benchmarks
Zhipu AI has released its new model, GLM-5.3, which is specifically designed for coding and security tasks. The model significantly improved performance on benchmarks, more than doubling its score on SWE-Marathon and qu…
-
xAI unveils Grok 4.5, targeting programmers and office tasks
xAI has launched Grok 4.5, its most advanced model to date, specifically designed for programming, agentic tasks, and knowledge work. Trained with Cursor on NVIDIA GB300 GPUs, Grok 4.5 shows strong performance in coding…
-
AI models struggle with marathon tasks, revealing benchmark limitations · 1 source tracked
New benchmarks reveal a significant gap between AI models' performance on short, single-session tasks and their ability to handle long, multi-hour operations. While models like GLM-5.2 and GPT-5.5 excel on benchmarks li…
-
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…
-
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…