2026
Mistral Ships Timestral Open Weights
Mistral's first time-series foundation models top GIFT-Eval, TIME and fev-bench, and run on a CPU
Read updateDit Model Radar
Model releases and changes, in one place.
Mistral's first time-series foundation models top GIFT-Eval, TIME and fev-bench, and run on a CPU
Read updatePlus: Grok Voice 3.0 Tops Speech Rankings, Liquid AI Opens 24B Local Model
Read updatePlus: Meta Ships First Frontier Model, Anthropic Targets Claude Copycats
Read updateIt out-scores Opus 5.5 on long-horizon coding by Meta's numbers, but trails it on the independent index
Read updateMistral's new 4.4B streaming speech-to-text model targets Arabic dialects and mid-sentence language switching
Read updateA fastText-based classifier that sorts Arabic script into 51 language and dialect classes
Read update