OpenAI
GPT-6.2 Astra
Total Context
1.1M
Max Output
128K
Released
N/A
gpt-5.4-nanoGPT-5.4 Nano is the lowest-cost, smallest GPT-5.4 option for high-volume simple tasks. OpenAI’s model guidance positions nano models around latency and cost efficiency rather than maximum reasoning depth. It should be used in descriptions for classification, routing, extraction, lightweight generation, and other predictable workflows.
Context Window
400K tokens
Maximum Output
128K tokens
Release Date
Mar 17, 2026
Modalities
| Input Price | Output Price | Cache Read |
|---|---|---|
| $0.2/M | $1.25/M | $0.02/M |
Reasoning
Tool calling
Temperature parameter
Attachments
Knowledge Base
Endpoint Protocols
GPT-5.4 nano prioritizes speed and processing economy for focused, high-volume text and image workloads.
Is optimized for simple, well-scoped tasks where rapid responses and economical execution are primary requirements.
Supports classification, extraction and ranking tasks that map large input sets into consistent labels, fields or ordered results.
Accepts up to 400,000 context tokens and image inputs, allowing focused processing across sizeable document collections.
GPT-5.4 nano is intended for large-scale classification, information extraction and ranking workloads with predictable output requirements.
Assigns predefined categories to high volumes of messages, documents or images for routing, analytics and moderation pipelines.
Extracts named fields from text, forms or document images and prepares consistent records for downstream systems.
Scores and orders search results, recommendations or support items according to explicit relevance criteria.
curl 'https://us-api.tokenhub.com/v1/messages' \
-X 'POST' \
-H "Authorization: Bearer $TOKENHUB_API_KEY"| Index score | ||
|---|---|---|
| Artificial Analysis Intelligence Index | Artificial Analysis broad capability aggregate | 38.2 |
| Artificial Analysis Coding Index | Artificial Analysis software task aggregate | 43.9 |
| Knowledge & Reasoning | ||
| GPQA | Advanced science problem solving | 81.7% |
| HLE | Broad expert-level exam set | 26.5% |
| Coding & Engineering | ||
| SciCode | Scientific coding challenges | 46.9% |
| Terminal-Bench Hard | Hard terminal task execution | 42.4% |
| Instruction Following & Agent Tasks | ||
| IFBench | Prompt constraint adherence | 75.9% |
| AA-LCR | Long-context reasoning | 66% |
| τ²-Bench | Agent workflow tasks | 76.0% |
Metrics sourced from Artificial Analysis
GPT-5.4 Nano: capabilities, use cases, limits, and TokenHub guidance.
GPT-5.4 Nano is a OpenAI model for simple high-volume tasks at the lowest GPT-5.4-class cost.
Best for simple extraction and classification, structured data generation and high-volume requests, especially when low-cost scale is the priority.
Key strength: the lowest-cost GPT-5.4-class option for simple high-volume work.
It is optimized for simple work and is not the best choice for difficult reasoning or large engineering tasks. For maximum answer quality, consider GPT-5.4 Mini.
Use the exact ID shown by TokenHub; follow your account docs and verify current features.
Use one API key to access GPT-5.4 Nano and more AI models through TokenHub.
Media and Discussions
Selected public videos and posts related to this model.
X (Twitter)
OpenAI ·
GPT-5.4 Thinking and GPT-5.4 Pro are rolling out now in ChatGPT. GPT-5.4 is also now available in the API and Codex. GPT-5.4 brings our advances in reasoning, coding, and agentic workflows into one frontier model. https://t.co/1hy6xXLAmJ
https://pbs.twimg.com/media/HCqogsobgAAAo2U.pngView originalOpenAI ·
GPT-5.4 is our most factual and efficient model: fewer tokens, faster speed. In ChatGPT, GPT-5.4 Thinking has improved deep web research, better context retention when it thinks for longer—and oh—you can now interrupt the model and add instructions or adjust its direction https://t.co/a0ayi6aOYV
https://pbs.twimg.com/media/HCqrhVeaUAYk72Y.jpgView originalOpenAI ·
GPT-5.4 mini is available today in ChatGPT, Codex, and the API. Optimized for coding, computer use, multimodal understanding, and subagents. And it’s 2x faster than GPT-5 mini. https://openai.com/index/introducing-gpt-5-4-mini-and-nano/ https://t.co/sirArgn37L
https://pbs.twimg.com/media/HDoN4PhasAAinj_.pngView originalReddit
GPT-5.4 Nano is genuinely impressive, how’s your experience?
Apprehensive_Fact710 ·
Read full post
I’ve been using GPT-5.4 Nano and I’m honestly blown away by how well it performs for being a smaller model. The speed feels great, and the output quality has been consistently strong for tasks I normally use larger models for. What I’m curious about: What kinds of prompts/workflows are you getting the best results with? How does it compare to models you were using before (quality, latency, reliability)? Any “best practices” you’ve found, prompt style, system instructions, or tool usage, that really improve results? Would love to hear your experience and any tips.
Introducing GPT-5.4 mini and nano
dayanruben ·
View originalGPT-5.4 nano for 0.25 premium request
Mindless-Push-1901 ·
Spotted in the documentation here : Supported AI models in GitHub Copilot - GitHub Docs (https://docs.github.com/en/copilot/reference/ai-models/supported-models) For everyone but free plan.
Supported AI models in GitHub Copilot - GitHub DocsView originalYouTube
AI Tanks: Mercury 2 vs GPT 5.4 Nano - Speed Matters!
Late Nights with AI
View originalOpenAI Announce GPT-5.4 Mini and GPT-5.4 Nano - High Level Intelligence At A Low Level Price.
Daily AI Roundup
View originalGPT-5.4 Nano: 20 Cents Per Million Tokens — 50x Cheaper Than GPT-5.4 #shorts #openai
Jaanus
View original