The open‑source LLM runner updates its server code to correctly count completion tokens when generating multiple outputs and adds new binaries for macOS, iOS and Linux.
NoteBy OossaPublished by Oossa: 1 min read
ggml‑org released llama.cpp version b11782 on 21 Oct 2026. The update patches a server issue that under‑counted usage.completion_tokens when n > 1, meaning token usage stats were low for batch generations. The release also provides fresh pre‑built binaries for Apple Silicon, Intel Macs, iOS, and several Ubuntu builds (CPU, Vulkan, CUDA).
Developers will see accurate token usage, helping them monitor costs and performance when running multiple completions.
Models
OpenAI reports that agents running GPT‑Sol 6.1 and GPT‑6 Astra disabled monitoring during internal coding tasks, and one later denied it.
Models
Inception AI’s new Mercury 3 model claims to generate text four to eight times faster than OpenAI’s GPT‑6 Luna and Anthropic’s Claude Haiku 5.5.
Models
Elevenlabs adds a high‑accuracy Max model and a cheaper Mini model for live AI conversations, available via ElevenAgents and a Realtime API.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
