Unverified50% confidenceBenchmarkExact time
Mistral released the Timestral time-series foundation model family on October 15, covering four sizes of 5M, 28M, 123M and 209M parameters; the largest model ranks first among pretrained zero-shot models on GIFT-Eval, TIME and fev-bench, with two-thirds the parameters of the previous best and without requiring ensembles.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/20/2026
First Seen
Valid until: 11/3/2026
Sources
Related Entities
Related Claims
UnverifiedThe 209M-parameter Timestral leads in the zero-shot model category on three benchmarks: GIFT-Eval, TIME and fev-bench.71% similarUnverifiedThe largest Timestral version (209M parameters) ranks first on three major public benchmarks.66% similarUnverifiedMistral 7B证明精心训练的7B模型可以匹敌早期的13B甚至30B模型63% similarUnverifiedMistral AI的7B模型在发布时性能超越了两倍参数量的Llama 2 13B61% similarUnverifiedTimestral supports 21-quantile probability forecasting and a context window of up to 32,768 steps.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074135API
curl https://kongchang.com/api/v1/knowledge/claims/1074135MCP
get_claim(id=1074135)