Unverified50% confidenceFactExact time
Mistral Large 3 has a total parameter count of 420 billion, and the decoding phase only needs to bear the compute cost and latency of 68 billion active parameters.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn a 3D generation prompt comparison, Mistral Large 4 cost a total of $0.87 across three scenarios, while Opus cost $13.83; Mistral disables reasoning by default, and with reasoning enabled the three scenarios took approximately 66 minutes.72% similarUnverified通义零码的MoE架构激活比例约为7.3%（350亿/4800亿），推理成本接近一个350亿参数的密集模型68% similarUnverifiedMixtral 8x7B拥有约470亿总参数，但每次推理仅激活约130亿参数，推理成本接近13B稠密模型66% similarUnverifiedFor large enterprises processing more than 500 million tokens per day, using an open-weight Mistral Large 3 local deployment can reduce the effective cost per million tokens to below $0.35, saving more than 80% compared with closed-source competitor APIs.64% similarPartially VerifiedMistral Large 4 is priced at approximately $0.68 per million input tokens and $2.09 per million output tokens, with a 50% discount for the first two weeks.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1008180API
curl https://kongchang.com/api/v1/knowledge/claims/1008180MCP
get_claim(id=1008180)