GPT-5.4 Mini: pricing, performance, and how to route requests

Route requests to
GPT-5.4 Mini
with Merge Gateway

Apply your own routing policies, reduce token costs automatically, and see every routing decision in real time with Merge Gateway.

What GPT-5.4 Mini costs to run

| Vendor | Input / 1M tokens | Output / 1M tokens | Zero data retention | | --- | ---: | ---: | --- | | OpenAI | $0.7500 | $4.50 | Yes |

Test GPT-5.4 Mini
with Gateway’s Simulator

See a prompt's output, token spend, latency, and more with GPT-5.4 Mini.

Route requests to GPT-5.4 Mini in minutes

To get started in seconds, add our Gateway Implementation skill to your project, or pick your preferred SDK below. Check out our other quick start skills here.

Install the Merge Gateway SDK

Python

1$ pip install merge-gateway-sdk

Send a request

Python

1from merge_gateway import MergeGateway
2
3client = MergeGateway(api_key="YOUR_API_KEY")
4
5response = client.responses.create(
6    model="openai/gpt-5.2",
7    input=[
8        {"type": "message", "role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
9        {"type": "message", "role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
10    ],
11)
12
13print(response.output[0].content[0].text)

Try a diffrent model

Swap the model string to route to a different provider. No other code changes needed.

Anthropic

1response = client.responses.create(
2    model="anthropic/claude-sonnet-4-20250514",
3    input=[
4        {"type": "message", "role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
5        {"type": "message", "role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
6    ],
7)

Point to Gateway

Python

1from openai import OpenAI
2
3client = OpenAI(
4    api_key="YOUR_API_KEY",
5    base_url="https://api-gateway.merge.dev/v1/openai",
6)

Send a request

Use the standard chat.completions.create method. No provider prefix needed on the model name.

Python

1response = client.chat.completions.create(
2    model="gpt-5.2",
3    messages=[
4        {"role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
5        {"role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
6    ],
7)
8
9print(response.choices[0].message.content)

Install packages

1npm install merge-gateway-ai-sdk-provider ai

Create the provider

TypeScript

1import { createMergeGateway } from "merge-gateway-ai-sdk-provider";
2
3const gateway = createMergeGateway({
4  apiKey: "YOUR_API_KEY",
5});

Send a request

Use generateText to send a request. Model names use the provider/model format.

TypeScript

1import { generateText } from "ai";
2
3const { text } = await generateText({
4  model: gateway("openai/gpt-4o"),
5  prompt: "Explain the concept of recursion in programming with a simple set of examples.",
6});
7
8console.log(text);

If you already have @ai-sdk/openai installed, point it at Gateway with a base URL change:

TypeScript

1import { createOpenAI } from "@ai-sdk/openai";
2
3const gateway = createOpenAI({
4  apiKey: "YOUR_API_KEY",
5  baseURL: "https://api-gateway.merge.dev/v1/ai-sdk",
6});
7
8// All generateText/streamText calls work unchanged

Install the Merge Gateway SDK

Anthropic SDK

1from anthropic import Anthropic
2
3client = Anthropic(
4    api_key="YOUR_API_KEY",
5    base_url="https://api-gateway.merge.dev/v1/anthropic",
6)
7
8message = client.messages.create(
9    model="claude-sonnet-4-20250514",
10    max_tokens=1024,
11    messages=[
12        {"role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
13    ],
14)
15
16print(message.content[0].text)

Explore other models available in Merge Gateway

Amazon Nova 2 Lite

Amazon Nova 2 Sonic

Amazon Nova Lite

Amazon Nova Micro

Amazon Nova Premier

Amazon Nova Pro

Bland Speech v3

Claude 3.7 Sonnet

Claude 4 Opus 20250514

Claude Fable 5

Claude Fable 5.1

Claude Haiku 4.5 (20251001)

Claude Haiku 5.5

Claude Opus 4.1 (20250805)

Claude Opus 4 (20250514)

Claude Opus 4.5 (20251101)

Claude Opus 4.6

Claude Opus 4.7

GPT-5.4 Mini FAQ

Have more questions about GPT-5.4 Mini? We've answered a few more below. Please note that this information was written in June, 2026 and is subject to change.

Heading

What other models does OpenAI offer?

OpenAI's lineup spans cost-optimized, general-purpose, and frontier reasoning tiers. Here are some other models OpenAI supports:

GPT-5.4 Nano: GPT-5.4 Nano is the cost-optimized sibling in the GPT-5.4 series, priced at $0.20 input and $1.25 output per million tokens. It scores 44 on the Intelligence Index and delivers a faster time to first token (6.33s) than GPT-5.4 Mini, making it the right choice when cost and speed are the primary constraints

GPT-4.1 Mini: GPT-4.1 Mini is a non-reasoning model at $0.40 input and $1.60 output per million tokens with a 1M token context window. It is faster to first token and cheaper than GPT-5.4 Mini, suited for tasks where reasoning traces are not required

GPT-5: GPT-5 is the standard mid-range model in the GPT-5 series with a 400k token context window, occupying the tier below the GPT-5.4 series

GPT-5.4: GPT-5.4 is the flagship in the GPT-5.4 series, scoring 57 on the Intelligence Index at $2.50 input and $15.00 output per million tokens. It delivers significantly higher capability than GPT-5.4 Mini at a higher price

GPT-5.5: GPT-5.5 is OpenAI's current highest-capability model at $5.00 input and $30.00 output per million tokens, ranked #2 on the Artificial Analysis Intelligence Index with a score of 60

How does GPT-5.4 Mini differ from OpenAI's other models?

GPT-5.4 Mini is OpenAI's mid-tier reasoning model in the GPT-5.4 series, combining meaningful intelligence with affordable pricing and official API support.

Pricing: GPT-5.4 Mini is priced at $0.75 per million input tokens and $4.50 per million output tokens. This is officially confirmed on OpenAI's pricing page, making it a reliable choice for production budgeting

Intelligence ranking: GPT-5.4 Mini scores 49 on the Artificial Analysis Intelligence Index, placing it #27 out of 150 evaluated models. This is higher than GPT-5 Mini at 41 and GPT-5.4 Nano at 44, while staying below GPT-5.4 at 57

Speed: Output speed is 164.4 tokens per second, one of the faster reasoning models. Time to first token is 12.81 seconds, substantially faster than the GPT-5 Mini (99.53s) it replaces for most workloads

Context window: GPT-5.4 Mini supports 400,000 input tokens, consistent with the rest of the GPT-5.4 series

Knowledge cutoff: GPT-5.4 Mini has a knowledge cutoff of August 2025, more than a year newer than GPT-5 Mini's May 2024 cutoff

Capabilities: Accepts text and image inputs, produces text output. Reasoning model released March 17, 2026

GPT-5.4 Mini is the right choice for teams that need reasoning-capable output with official pricing support, a reasonably fast TTFT, and intelligence meaningfully above the Nano tier.

What models should I consider using alongside GPT-5.4 Mini?

No single model is optimal for every task. Here are models worth pairing with GPT-5.4 Mini depending on what your product needs:

GPT-5.4 Nano: For the majority of requests that don't require GPT-5.4 Mini's intelligence tier, route to GPT-5.4 Nano at $0.20 input per million tokens. The Nano's TTFT is also faster at 6.33 seconds, making it viable for near-real-time batch workloads

GPT-5.4: When a task exceeds GPT-5.4 Mini's capability, escalate to GPT-5.4 at $2.50 input per million tokens for a 16-point Intelligence Index improvement. Use it selectively for ceiling tasks

GPT-4.1 Mini: For workloads that don't benefit from reasoning traces and need a larger context window, GPT-4.1 Mini's 1M token context at $0.40 input is a cost-efficient alternative

Claude Sonnet 4.6: For cross-provider diversity at a similar intelligence tier, Claude Sonnet 4.6 scores 44 on the Intelligence Index with a 1M token context window and competitive pricing from Anthropic

Mistral Small: For European data-residency workloads or cost-sensitive use cases where Mistral's infrastructure is preferred, Mistral Small provides a capable lightweight alternative in a similar price tier

What are the challenges of using GPT-5.4 Mini in my product?

Like any production LLM, GPT-5.4 Mini comes with tradeoffs worth planning for:

High time to first token for interactive use: At 12.81 seconds TTFT, GPT-5.4 Mini is too slow for synchronous user-facing interfaces. It works well for batch and async pipelines but cannot serve real-time chat or streaming UI patterns

Output verbosity: GPT-5.4 Mini generated 240 million output tokens during evaluation, which is among the higher verbosity levels in the dataset. At $4.50 per million output tokens, response length has a direct and significant cost impact

Smaller context window than alternatives: The 400,000 token context window is smaller than the 1M context available on GPT-4.1 Mini and GPT-4.1 Nano. Applications with very long documents, large codebases, or extended multi-turn histories may need a different model

Cost at scale: At $0.75 input and $4.50 output per million tokens, GPT-5.4 Mini is not a budget model. Combined with its verbosity, cost at production volume can be higher than initial estimates suggest

Provider dependency: Running exclusively on OpenAI creates fragility from outages, rate limits, or future model deprecations

Why should I use Merge Gateway to route LLM requests with GPT-5.4 Mini and every other model?

Using GPT-5.4 Mini through Merge Gateway gives you access to the model itself and the infrastructure layer around it:

One API, every provider: Access GPT-5.4 Mini and every other major LLM through a single endpoint and API key. Change providers by swapping the model string, with no application code changes required

Intelligent routing and automatic failover: Merge routes around OpenAI outages automatically. Routing policies based on cost, latency, or quality can reduce spend by routing simpler tasks to GPT-5.4 Nano or GPT-4.1 Mini without code changes

Cost governance: Set hard or soft project budgets so GPT-5.4 Mini spend stays within plan. Given its output verbosity, budget caps help prevent cost overruns. Every request is attributed to a model, project, and tag in a unified billing dashboard

Build Your Own Router: Score models against your own benchmark weights to determine when GPT-5.4 Mini is worth the cost versus GPT-5.4 Nano. Every routing decision includes a plain-language explanation of why a model was chosen

Security and compliance controls: Apply DLP rules and prompt injection protection before every request reaches OpenAI. Enforce per-project model and region policies without adding that logic to your application

How can I start using Merge Gateway to route requests with GPT-5.4 Mini?

Getting GPT-5.4 Mini running through Merge Gateway takes a few minutes:

1. Create an account and get your API key from the dashboard.

2. Install the Merge Gateway SDK: run pip install merge-gateway-sdk (Python) or npm install merge-gateway-sdk (Node). Alternatively, if you're already using the OpenAI SDK, set base_url = "https://api-gateway.merge.dev/v1/openai" and your existing code works as-is.

3. Make your first request using the provider/model format. For GPT-5.4 Mini, the model string is openai/gpt-5.4-mini. Swap the model string to route to any other provider without changing anything else.

4. Configure a routing policy in the dashboard to set failover behavior, cost limits, and optimization strategy. A practical starting point: use GPT-5.4 Nano as the default and escalate to GPT-5.4 Mini for tasks that need a higher intelligence tier, with GPT-5.4 as a further escalation option.

Full setup instructions and SDK references are in the Merge Gateway docs.