OpenCode Monthly Active Users Rose 28% to 22M in October
OpenCode's monthly active users grew 28% in October to 22 million, the largest jump since March, according to a post by jayair retweeted by the project.
Global AI Digest
-
-
Hyra 2.0 Trains Models From Scratch, Beating ProteinMPNN and qLDPC Baselines
Tencent's Hyra 2.0 research agent trained models from scratch, including a 1.96M-parameter inverse folding network running 57x faster than ProteinMPNN and a qLD…
-
Mistral Large 4 Writes a Quantum Chemistry Solver From Scratch
Mistral says its Mistral Large 4 model, nicknamed Le Chonk, wrote a quantum chemistry SCF solver from scratch, validated its output, and built a 3D visualizatio…
-
Tencent Hy Introduces Hyra 2.0 Research Agent
Tencent Hy announced Hyra 2.0, the next version of its Hunyuan Research Agent, which shifts from solving problems to building reusable tools, models, and workfl…
-
AI-Infra-Guard v4.6.8 Splits Scanners into Standalone pip Packages
Tencent's open-source AI-Infra-Guard v4.6.8 now ships each scanner as its own pip package, so users can install and run a single CLI without a web service.
-
OpenBMB Releases MiniCPM-OCR, a 1B-Parameter Document Parsing Model
OpenBMB released MiniCPM-OCR, a 1B-parameter OCR model for document parsing, and a Hugging Face Space now lets users compare its output against Baidu's Unlimite…
-
Command Code Weekly Recap: New Models, WeChat Pay, and /compact
Command Code's weekly recap adds DeepSeek V4.1 Flash, GLM 5.3 Flash, Ling 3.1 Flash and GPT-6.1 Terra, supports WeChat Pay, and demos the /compact command.
-
LiteLLM Introduces Unified Cost Accounting Engine
LiteLLM launched a Unified Cost Accounting Engine that provides a single source of truth for budgets and reports, dedicated workers for heavy analytics, and vis…
-
MiniMax M3.1 Flash Gains Community Quantization Support for Local Deployment
MiniMax highlights community efforts to quantize its 428B-parameter M3.1 Flash model, including a 2-bit Dynamic GGUF build at 143.9 GB and other low-bit recipes…
-
OpenCode's GitHub Triage Runs as a Long-Running Session That Spawns Specialized Subagents
Dax describes how OpenCode's GitHub triage system works from a long-running OpenCode session, delegating tasks to subagents that have handled similar work befor…
