Enjoying AI Stats?Support us
US
OpenAI

GPT-4o-mini

OpenAI
Overview

A smaller version of the GPT-4o series model, designed to make AI more accessible and affordable. It excels in basic text based tasks, outperforming previous models like GPT-3.5 Turbo, offering superior performance, for an accessible price.

Quick Links
Key Metrics
Max Input
128,000
Tokens
Max Output
16,384
Tokens
Throughput
-
tok/s
Latency
-
ms
Input Price
$0.15
Per 1M Tokens
Cached Input Price
$0.075
Per 1M Tokens
Output Price
$0.6
Per 1M Tokens
Blended Price
$0.488
Per 1M Tokens
Model Information
Release Details

Released

18 Jul 2024

Knowledge Cutoff

-

License

Proprietary

Model Architecture

Parameters

-

Training Data

-

Context Window

Input Context Length

128,000 tokens

Output Context Length

16,384 tokens

Key Features
Web Access
Unknown

Real-time access to current web information

Multimodal
Yes

Ability to process multiple data types (text, images, etc.)

Reasoning
No

Advanced logical and deductive reasoning capabilities

Fine-Tunable
Unknown

Can be customized for specific use cases

Model Release & Updates
18 Jul 2024
Model Released
Model first made available to the public
13 Jun 2023
Previous version
14 Mar 2023
Previous version
Benchmarks & Performance Comparison
AI Stats Score
1302.25
#149/191
Confabulations
37.21
Lower is better
#36/37
Elimation Game
4.60
#17/26
GPQA Diamond
40.20%
#72/94
LMArena Text
1312.00
#41/76
NYT Connections
9.90%
#41/43
SimpleBench
10.70%
#31/31
SWE-Bench
9.00%
#12/12
Thematic Generalisation
2.30
Lower is better
#33/34