MiniMax M3.1 Flash Now Free in Cline, Aimed at Long Agentic Coding Runs
MiniMax M3.1 Flash is available free in Cline from day one, positioned for long agentic coding runs where costs accumulate quickly, with roughly 25x lower cost…
Global AI Digest
-
-
MiniMax M3.1 Flash Revealed as OpenRouter's 'Space Bunny' Stealth Model, Free for 72 Hours
MiniMax confirmed that the stealth model 'Space Bunny' on OpenRouter was an early build of M3.1 Flash, which topped the platform for 10 days and processed 56.7T…
-
MiniMax M3.1 Flash Tops OpenCode for 12 Days, Free for 3 Days
MiniMax revealed that the anonymous "Space Bunny" model was M3.1 Flash, which held #1 on OpenCode for 12 consecutive days and served 106T tokens; the finished v…
-
MiniMax M3.1 Flash Lands on Ollama Cloud With US Hosting and Zero Data Retention
MiniMax M3.1 Flash is now available on Ollama Cloud, hosted in the US with zero data retention, a 1M-token context window, image understanding, and support for…
-
Command Code Adds Customizable /statusline for Model, Usage and Cost Info
Command Code launched a /statusline command that shows model, usage, reasoning level and costs below the prompt, with options to reorder items and toggle indica…
-
Meta Publishes Technical Report on Muse Glimmer, a 30B Open-Weight Model for Agentic Tasks on Consumer Hardware
Meta released a technical report on Muse Glimmer, a 30B open-weight model designed for agentic tasks, running at practical speeds on a single consumer GPU at 4-…
-
ChatGPT Slides Adds Real-Time Collaboration and On-Slide Editing
ChatGPT Slides now supports real-time co-editing, comments, and direct on-slide editing, with export to PowerPoint or Google Slides, available on Pro, Business,…
-
Command Code Adds Customizable Statusline to Its Coding Agent
Command Code launched a /statusline command that lets users customize the line below the prompt, showing model, usage, reasoning level, and costs with reorderab…
-
MiniMax M3.1 Flash Now Available on Ollama Cloud
MiniMax M3.1 Flash is available on Ollama Cloud with US hosting, zero data retention, image understanding, and a 1M-token context window, usable via Claude Code…
-
Meta Details Prometheus Supercluster Powering Muse Models
Meta shared engineering details of Prometheus, its gigawatt-scale supercluster in New Albany, Ohio, used to train Muse models including Muse Wonder, reporting 8…
