MistralOpen SourceDec 28, 2023

Mistral: Mistral 7B Instruct v0.2

A high-performing, industry-standard 7.3B parameter model, with optimizations for speed and context length. An improved version of Mistral 7B Instruct, with the following changes: - 32k context window (vs 8k context in v0.1) - Rope-theta = 1e6 - No Sliding-Window Attention

Context Window
33K
tokens
Max Output
tokens
Released
Dec 28, 2023
Arena Rank

Capabilities

👁Vision
🧠Reasoning
🔧Tool Calling
Prompt Caching
🖥Computer Use
🎨Image Generation

Supported Parameters

Frequency Penalty
Reduce repetition
Logit Bias
Adjust token weights
Max Tokens
Output length limit
min_p
Presence Penalty
Encourage new topics
Repetition Penalty
Penalize repeated tokens
Stop Sequences
Custom stop tokens
Temperature
Controls randomness
Top K
Top-K sampling
Top P
Nucleus sampling

Pricing Comparison

RouterInput / 1MOutput / 1MCached Input / 1M
OpenRouter$0.20$0.20
Martian$0.20$0.20

Benchmarks

Open LLM Leaderboard
AverageOpen LLM Leaderboard
18.51/100
IFEvalOpen LLM Leaderboard
54.96/100
BBHOpen LLM Leaderboard
22.91/100
MATH Lvl 5Open LLM Leaderboard
3.02/100
GPQAOpen LLM Leaderboard
3.47/100
MUSROpen LLM Leaderboard
7.61/100
MMLU-PROOpen LLM Leaderboard
19.08/100

Model IDs

OpenRoutermistralai/mistral-7b-instruct-v0.2
Compare with another model

Related Models