modelgrep
O

OpenAI: GPT-5 Codex

openai/gpt-5-codex

24th smartest of 178ReasoningToolsJSONVision
Use via OpenRouter ↗
Intelligence
44.6
24th of 178
Design Elo
1140
mobileapps
Speed
43
194th fastest
Latency
3.6s
first token
Input price
$1.25
225th cheapest
Context
400K
128K max out

How it compares

Smarter than87%
of all ranked models
Faster than34%
of all ranked models
Cheaper than24%
of all ranked models

Overview

GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows. It is designed for both interactive development sessions and long, independent execution of complex engineering tasks....

Benchmarks

independent · via OpenRouter
Artificial Analysis83th percentile
Intelligence Index
44.6
Coding Index
38.9
Agentic Index
52.7
GPQA Diamond
84%
Humanity's Last Exam
26%
SciCode
41%
Tau²-Bench (agentic)
87%
Design Arena · Elo845 tournaments
mobileapps
1140

Providers & pricing (1)

ProviderIn $/MOut $/MUptime
OpenAIcache$1.25$10.00

Specifications

Context window400K
Max output128K
Knowledge cutoffSep 2024
Input modalitiestext, image
Output modalitiestext
Prompt cachingSupported
Cache read price$0.125/M
ModeratedYes

GPT-5 Codex FAQ

How much does GPT-5 Codex cost?

GPT-5 Codex costs $1.25 per million input tokens and $10.00 per million output tokens via OpenRouter, making it 225th cheapest of 298 paid models.

How smart is GPT-5 Codex?

GPT-5 Codex scores 44.6 on the Artificial Analysis Intelligence Index, ranking 24th of 178 benchmarked models, with a GPQA Diamond score of 84%.

How fast is GPT-5 Codex?

GPT-5 Codex generates around 43 tokens per second with 3.6s time-to-first-token (p50), the 194th fastest tracked model.

What is GPT-5 Codex's context window?

GPT-5 Codex supports a 400K-token context window and can output up to 128K tokens. It accepts text, image input.

Compare head-to-head