Mistral: Ministral 8B

Text input Text output
Author's Description

Ministral 8B is an 8B parameter model featuring a unique interleaved sliding-window attention pattern for faster, memory-efficient inference. Designed for edge use cases, it supports up to 128k context length and excels in knowledge and reasoning tasks. It outperforms peers in the sub-10B category, making it perfect for low-latency, privacy-first applications.

Key Specifications
Cost
$$
Context
128K
Parameters
8B
Released
Oct 16, 2024
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Stop Presence Penalty Tool Choice Top P Temperature Seed Tools Structured Outputs Response Format Frequency Penalty Max Tokens
Features

This model supports the following features:

Tools Structured Outputs Response Format
Performance Summary

Mistral: Ministral 8B, created on October 16, 2024, is an 8B parameter model designed for edge use cases, featuring an interleaved sliding-window attention pattern for efficient inference and supporting up to 128k context length. The model performs among the fastest models, ranking in the 64th percentile for speed across six benchmarks. It consistently offers highly competitive pricing, placing in the 87th percentile. Furthermore, Ministral 8B demonstrates exceptional reliability, achieving a perfect 100th percentile, indicating minimal technical failures and consistent response delivery. In terms of performance across benchmark categories, Ministral 8B shows a mixed profile. It exhibits strong accuracy in Email Classification (96.0%) and General Knowledge (91.5%), though its percentile rankings in these areas (44th and 38th respectively) suggest a competitive but not leading position. Its Ethics performance is also high at 96.5% accuracy, albeit at the 33rd percentile. A notable strength is its cost-efficiency, consistently ranking in high percentiles across all benchmarks, particularly in Reasoning (94th percentile) and Ethics (91st percentile). However, the model shows significant weaknesses in Reasoning (26.0% accuracy, 16th percentile) and Instruction Following (40.4% accuracy, 40th percentile), indicating areas for improvement in complex logical deduction and multi-step directive execution. Coding performance is moderate at 77.0% accuracy (45th percentile).

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.1
Completion $0.1

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Mistral
Mistral | mistralai/ministral-8b 128K $0.1 / 1M tokens $0.1 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Free Executions Accuracy Cost Duration
Other Models by mistralai