What other models does OpenAI offer?

OpenAI maintains a tiered lineup spanning cost-optimized, general-purpose, and frontier reasoning models. Here are some other models OpenAI supports:

• GPT-4o: GPT-4o is OpenAI's general-purpose multimodal flagship before the GPT-5 series. It accepts text and images, produces text output, and runs at substantially lower cost than GPT-5.5, making it a strong default for latency-sensitive or high-volume tasks where frontier reasoning is not required

• GPT-4.1 / GPT-4.1 mini / GPT-4.1 nano: This is a cost-efficient tier positioned below GPT-4o for applications that prioritize throughput and price over intelligence. The nano variant is suited for classification, routing, and lightweight extraction tasks at very low per-token cost

• GPT-5.2: GPT-5.2 is an earlier GPT-5 series release that is less capable than GPT-5.5 on frontier benchmarks but meaningfully lower in cost, making it a reasonable choice when GPT-5.5's intelligence gains are not worth the price premium

• GPT-5-mini: GPT-5-mini is a smaller, faster version of the GPT-5 family optimized for cost and speed rather than maximum reasoning depth, suited to high-volume tasks that benefit from GPT-5 architecture without the flagship price

• o3 / o4-mini: These are dedicated reasoning models that produce chains of thought before answering and excel at math, logic, and multi-step problem solving. o4-mini trades some capability for significantly lower cost than full reasoning-tier models

How does GPT-5.5 differ from OpenAI's other models?

GPT-5.5 is OpenAI's highest-intelligence production model, sitting above the rest of the GPT-5 series on frontier benchmarks and priced accordingly.

• Pricing: GPT-5.5 costs $5.00 per million input tokens and $30.00 per million output tokens. GPT-5.2 and GPT-5-mini are meaningfully cheaper. For context, GPT-5.5 doubled the API price of its predecessor GPT-5.4

• Intelligence ranking: GPT-5.5 scores 60 on the Artificial Analysis Intelligence Index, placing it #2 of 150+ evaluated models. GPT-4o and the GPT-4.1 tier score substantially lower, reflecting the generational gap between the series

• Context window: GPT-5.5 supports approximately 1 million input tokens, compared to 128K for GPT-4.5 and smaller windows on most GPT-4.x variants. This makes it one of the few models suitable for very long document processing without chunking

• Latency: Time to first token averages 58.37 seconds, placing it at the slower end of all evaluated models. This is expected for a reasoning model with extended thinking, but rules it out for any real-time or interactive use case

• Output verbosity: GPT-5.5 generated approximately 75 million output tokens during Artificial Analysis evaluation, versus a median of 35 million across all models. At $30 per million output tokens, this verbosity materially inflates costs compared to what pricing alone suggests

• Benchmarks: Terminal-Bench 2.0: 82.7%, GDPval: 84.9%, SWE-Bench Pro: 58.6%. These results rank it at or near the top on software engineering and agentic task benchmarks

GPT-5.5 is the right choice when task difficulty genuinely requires frontier reasoning and output quality, and when latency is not a constraint. For most high-volume production workloads, a cheaper sibling handles the load at a fraction of the cost.

What models should I consider using alongside GPT-5.5?

No single model is optimal for every task. Here are models worth pairing with GPT-5.5 depending on what your product needs:

• Claude Opus 4.8 (Anthropic): Scores #1 on the Artificial Analysis Intelligence Index at 61, edging GPT-5.5 by one point. Route your most demanding reasoning tasks, especially those requiring nuanced judgment or extended analysis, to Opus 4.8 when maximum output quality is the only metric that matters

• GPT-5-mini (OpenAI): Use for any task in your pipeline that does not require frontier reasoning: classification, extraction, summarization, or formatting. Routing these requests away from GPT-5.5 significantly reduces costs given GPT-5.5's output pricing and verbosity

• Gemini 2.5 Flash (Google): Delivers a first-token response in approximately 0.36 seconds, compared to GPT-5.5's 58 seconds. Route latency-sensitive requests, live chat interfaces, or streaming UI interactions to Gemini 2.5 Flash where response speed matters more than deep reasoning

• o4-mini (OpenAI): A cost-optimized reasoning model for math, code, and logic tasks that benefit from chain-of-thought but do not require GPT-5.5's full capability. Suitable for medium-complexity reasoning at a lower price point within the same provider

• Kimi K2.6 (Moonshot AI): The highest-ranked open-weight model on the Intelligence Index at a score of 54. Consider for workloads where you want strong intelligence without vendor lock-in to a proprietary provider, particularly for batch processing where per-token cost is the primary constraint

What are the challenges of using GPT-5.5 in my product?

Like any production LLM, GPT-5.5 comes with tradeoffs worth planning for:

• Output verbosity inflates cost unpredictably: GPT-5.5 generates approximately twice the output tokens of the median LLM. At $30 per million output tokens, this verbosity compounds quickly. A task that looks affordable on input pricing can cost significantly more once actual output volume is measured in production

• Long-context pricing surcharges: Requests that exceed 272K input tokens are billed at 2x the standard input rate and 1.5x the output rate for the full session. Applications relying on GPT-5.5's million-token context window need to account for this pricing tier in cost modeling

• Not viable for real-time interfaces: A time-to-first-token of 58 seconds makes GPT-5.5 unsuitable for chat UIs, live completions, or any interaction where users are waiting for a response. This is a hard constraint, not one that routing alone solves

• Provider dependency: Running entirely on OpenAI means a provider outage, rate limit, or model deprecation directly impacts your product. GPT-5.4 was a direct predecessor to GPT-5.5, indicating the model series moves quickly, and API availability is not guaranteed indefinitely for any given version

• Cost at scale: At $5 input and $30 output per million tokens, GPT-5.5 is one of the most expensive production models available. Without project-level budgets and active routing controls, costs can scale unexpectedly as usage grows

Why should I use Merge Gateway to route LLM requests with GPT-5.5 and every other model?

Using GPT-5.5 through Merge Gateway gives you access to the model itself and the infrastructure layer around it:

• One API, every provider: Access GPT-5.5 and every other major LLM through a single endpoint and API key. Change providers by swapping the model string—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 40–60%, and given GPT-5.5's output pricing, routing even a fraction of lower-complexity requests to a cheaper model has an outsized impact

• Cost governance: Set hard or soft project budgets so GPT-5.5 spend stays within plan — particularly useful given its verbosity at $30 per million output tokens. Every request is attributed to a model, project, and tag in a unified billing dashboard across all providers

• Build Your Own Router: Define what "best" means for your traffic by selecting from curated ML benchmarks or adding your own eval scores. The router scores each available model against your weights and picks the winner per request, with a plain-language explanation of every decision

• 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-5.5?

Getting GPT-5.5 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-5.5, the model string is openai/gpt-5.5. 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 for GPT-5.5: name it as primary for complex tasks, with GPT-5-mini or o4-mini as fallback for when budgets are near limit or provider latency spikes.

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