Unverified50% confidenceReleaseExact time
Mistral launched Timestral, its first time-series foundation model family, with four sizes (5M, 28M, 123M, and 209M), running purely on CPU, with per-forecast latency of 10–140ms and support for a 32768-step context.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedOpen-source models with fewer than 7B parameters, such as Llama 3 and Mistral 7B, combined with 4-bit quantization, can run on a single consumer-grade GPU or a high-performance CPU.66% similarUnverifiedIn a pure-CPU environment, comparing llama.cpp build 9976 and build 11571 running the same Gemma quantised model, the new build's generation speed is about 9.1% faster than the old build (6.60 vs 6.05 tok/s).62% similarUnverified实测项目使用的芯片型号为STM32F103C8T6，系统时钟配置为72MHz，GPIO配置在PC13引脚62% similarUnverifiedThe inference process for LLaVA-7B processing a single image requires approximately 2,000–4,000 matrix multiplication operations, with an end-to-end latency of approximately 300–800 milliseconds on an A100 GPU.60% similarUnverifiedA 1.5-millisecond CPU inference means the detection module can be embedded into existing inference pipelines without additional GPU resources.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1052860API
curl https://kongchang.com/api/v1/knowledge/claims/1052860MCP
get_claim(id=1052860)