OpenAI
The full GPT-5.x hierarchy plus the o-series reasoning line — every tier on Nagent.
OpenAI's 2026 lineup is a hierarchy, not a single model. GPT-5.5 (the frontier flagship) hit 88.7% on SWE-bench and 92.4% MMLU at launch, with a 60% reduction in hallucinations versus GPT-4. The series fans out from gpt-5.5-pro for high-stakes legal and medical work down to gpt-5.4-nano for edge inference, plus the o-series (o3, o4-mini) for chain-of-thought reasoning, the Codex sub-models for software engineering, and gpt-realtime-1.5 for voice agents.
Models
GPT-5.5
Flagship88.7% SWE-bench, 92.4% MMLU — frontier general-purpose reasoning
Context Window
272K
Max Output
32K tokens
Input Price
$5.00 / 1M
Output Price
$30.00 / 1M
Input Types
Text, Images
Output Types
Text, Code, JSON
GPT-5.5 Pro
EnterpriseHigh-stakes legal, medical, and regulated domains
Context Window
272K
Max Output
32K tokens
Input Price
$30.00 / 1M
Output Price
$180.00 / 1M
Input Types
Text, Images
Output Types
Text, Code, JSON
GPT-5.4
DefaultDefault development standard — best speed/quality balance
Context Window
128K
Max Output
16K tokens
Input Price
$2.50 / 1M
Output Price
$15.00 / 1M
Input Types
Text, Images
Output Types
Text, Code
GPT-5.4 Mini
High-volume automation pipelines
Context Window
128K
Max Output
16K tokens
Input Price
$0.75 / 1M
Output Price
$4.50 / 1M
Input Types
Text, Images
Output Types
Text, Code
GPT-5.4 Nano
Edge tier — low-complexity tasks at scale
Context Window
128K
Max Output
8K tokens
Input Price
$0.20 / 1M
Output Price
$1.25 / 1M
Input Types
Text
Output Types
Text, Code
o4-mini
ReasoningChain-of-thought reasoning for STEM and logic
Context Window
128K
Max Output
100K tokens
Input Price
$1.10 / 1M
Output Price
$4.40 / 1M
Input Types
Text, Images
Output Types
Text, Code
o3 Deep Research
ResearchAcademic synthesis across multiple sources
Context Window
256K
Max Output
100K tokens
Input Price
$10.00 / 1M
Output Price
$40.00 / 1M
Input Types
Text, Images
Output Types
Text, Code
GPT-5.3 Codex
Specialised coding endpoint — outperforms general models on code metrics
Context Window
128K
Max Output
16K tokens
Input Price
$2.50 / 1M
Output Price
$15.00 / 1M
Input Types
Text, Code
Output Types
Code
GPT-realtime-1.5
VoiceInteractive voice agents with sub-second latency
Context Window
128K
Max Output
8K tokens
Input Price
$4.00 / 1M
Output Price
$16.00 / 1M
Input Types
Text, Audio
Output Types
Audio, Text
GPT-4o (legacy)
Legacy real-time voice & vision flagship
Context Window
128K
Max Output
16K tokens
Input Price
$2.50 / 1M
Output Price
$10.00 / 1M
Input Types
Text, Images, Audio
Output Types
Text, Code, JSON
Use Cases
High-Stakes Reasoning
Use GPT-5.5-Pro for legal review, medical documentation, or regulated workflows where the cost of a wrong answer dwarfs token cost.
High-Volume Production
GPT-5.4-mini and -nano handle millions of classification, extraction, and routing tasks per day at fractional cost.
Complex Reasoning Tasks
o3 and o4-mini bring chain-of-thought processing to scientific analysis, financial modelling, and multi-step deductions.
Voice & Multimodal
gpt-realtime-1.5 powers conversational voice agents; GPT-5.4 handles vision tasks like document parsing and chart analysis.
Why use OpenAI on Nagent?
Nagent adds enterprise orchestration, observability, and workflow automation on top of OpenAI's raw model capabilities.
Function-calling and parallel tool use natively supported in our agent definitions
Switch between GPT-5.5, o3, and Codex mid-workflow based on task type
JSON mode enforced at the platform level — no prompt hacking required
Built-in rate-limit handling and automatic failover between tiers
How to access OpenAI on Nagent
Open Agent Studio
Navigate to Agent Studio in your Nagent workspace.
Select OpenAI
Choose OpenAI and pick the right tier — flagship for hard problems, mini/nano for volume, o-series for reasoning.
Enable Tools
Toggle function-calling or JSON mode in the model settings and connect your data sources.
Common questions about OpenAI
Real buyer and developer questions, answered. Click any item to expand.
GPT-5.5 vs Claude Opus 4.7 — which should I pick?
Opus 4.7 wins on multi-file engineering and natural prose quality. GPT-5.5 wins on broad general-purpose reasoning, vision tasks, and is the practical default if you also need o-series chain-of-thought reasoning in the same workflow.
When should I use the o-series (o3, o4-mini) instead of GPT-5.5?
When the task needs explicit chain-of-thought — STEM problems, scientific reasoning, multi-step logic, financial modelling. For everyday agent loops (drafting, classification, tool-use) GPT-5.4 is faster and cheaper.
Does Nagent support OpenAI's Realtime API for voice agents?
Yes — gpt-realtime-1.5 is wired into our voice-agent skill alongside Cartesia and ElevenLabs. Pick OpenAI when you want a single provider for both reasoning and speech in the same agent.
Is GPT-5.5 Pro at $30/$180 per 1M tokens ever worth it?
Yes for high-stakes outputs — legal review, medical documentation, regulated workflows — where the cost of a wrong answer dwarfs token cost. Use it sparingly through Nagent's tier-aware router so cheaper models handle the routine subtasks.
Ready to use OpenAI inside your agents?
Get started in minutes — no API key management required.