Unverified50% confidenceFactExact time
Mixtral 8x7B, released by Mistral AI, is the first mainstream open-weight MoE model, with 46.7B total parameters and 12.9B active parameters.
1
Sources
50%
Confidence
Long-term
Relevance
7/23/2026
First Seen
Sources
Qwen3 235B A22B实测：2.4万亿参数开源模型超越Claude Opus
Related Claims
UnverifiedMistral AI has released Magistral Small version 1.1, an efficient reasoning model with 24 billion parameters.69% similarVerifiedMeta's LLaMA series, Mistral AI's Mistral/Mixtral series, Alibaba's Qwen series, and the DeepSeek series all offer open-source versions with parameter scales ranging from 1B to hundreds of billions.68% similarUnverifiedThis test featured 7 AI models: Claude Sonnet 4.6, DeepSeek V4 Flash, Kimi K2.6, GLM 5.1, MiniMax M2.7, Mimo V2.5 Pro, and Qwen 3.6 Plus.61% similarUnverifiedNew-generation small models such as Mistral 7B/8x7B, Qwen 2.5, and Gemma improve performance by enhancing the Transformer architecture (e.g., GQA, sliding window attention) and improving training data quality.59% similarUnverifiedTypical representatives of the MoE architecture include Mixtral 8x7B and DeepSeek-V2.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/600089API
curl https://kongchang.com/api/v1/knowledge/claims/600089MCP
get_claim(id=600089)