Mistral Small (Sep '24)
MistralMistral开源权重Apache 2.0 · 商用许可
描述
Mistral Small 4 is Mistral AI's hybrid small model that unifies instruct, reasoning, multimodal understanding, and coding capabilities in a single efficient system. It uses a 119B-parameter Mixture-of-Experts architecture with native image input, a 256k context window, and configurable reasoning depth for general chat, coding agents, and long-context reasoning workloads.
发布日期
2024-09-17
参数规模
119.0B
上下文长度
256K
支持模态
image, text
能力雷达图
20
general
14
coding
28
reasoning
28
science
24
agents
60
multimodal
排行榜排名
基准测试分数 (LLM Stats)
(LLM Stats (zeroeval))General
MMLU
80.7%自报
TriviaQA
80.3%自报
Arena Hard
58.3%自报
Instruction Following
IFBench
48.0%自报
Language
MMLU-Pro
78.0%自报
COLLIE
62.9%自报
Math
AIME 2025
83.8%自报
GSM8k
80.7%自报
MATH
46.0%自报
Reasoning
ARC-C
91.3%自报
GPQANYU + Cohere + Anthropic (2023)
71.2%自报
AA-LCR
71.2%自报
MBPP
0.70 / 100自报
AGIEval
65.8%自报
LiveCodeBench
63.6%自报
Vision
MMMU-Pro
60.0%自报
AA 评测指数
(Artificial Analysis)Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))56.3
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))52.9
Gpqa(NYU + Cohere + Anthropic (2023))38.1
Livecodebench(UC Berkeley + MIT + Cornell (2024))14.1
Aime(MAA (Mathematical Association of America))6.3
Intelligence Index(Artificial Analysis)5.8
Hle(Center for AI Safety + Scale AI (2025))4.3
LLM Stats 分类评分
(LLM Stats (zeroeval))Legal80
Math80
Finance80
Healthcare80
Language70
Long Context70
Physics70
Reasoning70
General70
Biology70
Chemistry70
Chat60
Multimodal60
Code60
Creativity60
Vision60
Writing60
Instruction Following50
定价
输入价格免费
输出价格免费
混合价格(3:1)免费
速度
Tokens/秒0.0
首Token延迟0.00s
首回答延迟0.00s
供应商价格排行
供应商价格排行
2 个供应商
最便宜: Mistral AI最贵: Helicone
供应商输入输出
1Mistral AI最便宜
$0
$0
2Helicone
$0.075
$0.2
比较该模型在不同 API 供应商之间的定价。