| [x] MageVL |
https://www.reddit.com/r/LocalLLaMA/comments/1v97f8d/microsoftmagevl_hugging_face_an_efficient/ |
tool, framework |
integrated |
MageVL |
Efficient vision-language model framework from Microsoft. |
| [x] BetterGPT-150M |
https://www.reddit.com/r/LocalLLaMA/comments/1v9oa1u/built_and_released_bettergpt150m_a_compact_150m/ |
tool |
integrated |
BetterGPT-150M |
A compact 150M parameter language model. |
| [x] LFM-2.5 Encoders |
https://huggingface.co/blog/LiquidAI/lfm2-5-encoders |
tool, provider |
integrated |
LFM-2.5 Encoders |
Liquid AI's latest encoder models. |
| [x] OLMoEarth |
https://huggingface.co/blog/allenai/olmoearth-infrastructure |
infrastructure, tool |
integrated |
OLMoEarth |
Infrastructure for Earth-specific language modeling by AllenAI. |
| [x] Gemini API Managed Agents |
https://blog.google/innovation-and-ai/technology/developers-tools/expanding-managed-agents-gemini-api-3-6-flash-hooks/ |
provider, tool |
integrated |
Gemini API Managed Agents |
Google's managed agent capabilities for Gemini API. |
| [x] Diagrid Catalyst |
https://thenewstack.io/diagrid-catalyst-agent-recovery/ |
tool, infrastructure |
integrated |
Diagrid Catalyst |
Agent recovery and orchestration platform. |
| [x] Grafana Assistant Data Source |
https://www.infoq.com/news/2026/07/grafana-assistant-data-source/ |
tool |
integrated |
Grafana Cloud |
AI-powered data source integration for Grafana. |
| [x] SWE-bench Multilingual |
https://www.reddit.com/r/LocalLLaMA/comments/1v93phk/swerebench_multilingual_update_go_java_python/ |
benchmark/eval |
integrated |
SWE-bench |
Benchmark for evaluating software engineering agents across multiple languages. |