01Overview02Models03Rankings04Compare05Quiz06Guide

Model

GPT-5.4 nano NO.15

OpenAI's cheapest small model for high-volume simple work.

CompareBack to the index

OpenAI · GPT · Fast · Closed weights

Specification

Fig.01 — GPT — OpenAI

USA

Context window
400Ktokens
Max output
128Ktokens
Input price
$0.20/ 1M
Output price
$1.25/ 1M
Cached input
$0.02/ 1M
Throughput
—
Class
Fast
Modalities
Text, Image, Pdf
Weights
Closed
Knowledge cutoff
Oct 2025
Released
Mar 2026

Overview

GPT-5.4 nano is OpenAI's cheapest and smallest current model, built for high-volume, simple, latency-sensitive tasks. It is ideal as a router target for classification, extraction, and lightweight generation where cost dominates. It is not meant for complex reasoning or agentic coding.

Strengths

  • Cheapest model in the lineup ($0.20/$1.25)
  • Very fast, low-latency output
  • 400K-token context despite tiny size
  • Multimodal text, image, and PDF input
  • Aggressive cache discount for repeated prompts

Trade-offs

  • Limited reasoning and coding depth
  • Not for complex or agentic tasks
  • Weaker instruction adherence on hard prompts

Fit

Best for

  • High-volume classification and extraction
  • Routing and intent detection
  • Cheap autocomplete and suggestions
  • Bulk tagging and moderation
  • Latency-critical lightweight calls

Not ideal for

  • Complex reasoning or coding
  • Agentic work
  • High-quality long-form writing

Capabilities

Ten axes, normalized 0–100 and scored the same way across the whole catalog — so a 78 here means what a 78 means anywhere else on the bench.

Fig.02 — Capability radar

0 — 100

Reasoning
66
Coding
68
Math
70
Writing
74
Knowledge
72
Speed
95
Agentic
64
Vision
70
Multilingual
78
Long Context
78

Benchmarks

Public results, with independent third-party runs marked. Bars normalize percentages against 100 and Elo ratings against a 1500 ceiling.

Independent index

Artificial Analysis Intelligence Index

Composite of ~9–10 independent evals · Artificial Analysis

38/ 100

BenchmarkResult
Artificial Analysis Intelligence IndexIndependentGeneral · Artificial Analysis
38%
MMLU-ProKnowledge · Vendor-reported
35.6%
LiveBenchIndependentGeneral · LiveBench
70.1%

Alternatives

Models in roughly the same class — the ones worth weighing against this record.

Compare side by side
01GPT-5.4 miniOpenAILow-cost previous-generation OpenAI workhorse for standard routing.02Gemini 3.1 Flash-LiteGoogle DeepMindCheapest Gemini, optimized for low-latency high-volume traffic.03Claude Haiku 4.5AnthropicFastest, cheapest Claude — snappy everyday tasks at scale.

Index

PreviousGPT-5.4 miniNextGemini 3.5 Pro
All models/EDUCATION