The lab that started the frontier AI race.
OpenAI created GPT-4o, GPT-5.4, and GPT-5.5 — the models powering ChatGPT and thousands of enterprise products. Their API is the most widely used in the industry.
Every OpenAI model in the directory, ranked by overall capability score.
GPT-5 is OpenAI's flagship multimodal model, superseding GPT-4o with significantly improved reasoning, instruction-following, and knowledge breadth. It handles text, images, and complex multi-step tasks with state-of-the-art performance across most benchmarks.
Per 1 million tokens. Updated when providers change prices.
| Model | Input / 1M | Output / 1M | Context | Speed |
|---|---|---|---|---|
| OpenAI: GPT-5 Balanced | $30.00/1M | $180.00/1M | 400K | Balanced |
| GPT-5.5 Premium | $5.00/1M | $30.00/1M | 1M | Balanced |
| GPT-5.4 Premium | $0.20/1M | $1.25/1M | 272K | Balanced |
| OpenAI: GPT-5 Image Premium | $10.00/1M | $10.00/1M | 400K | Balanced |
| OpenAI: GPT-5 Image Mini Balanced |
Head-to-head comparisons for the most-searched questions.
Newsletter
Pricing changes, new releases, and ranking shifts — straight to your inbox.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.
GPT-5.5 is OpenAI's most capable model in 2026 — it leads on Terminal-Bench (82.7%) and integrates with Codex for agentic workflows. GPT-5.4 is the only model that can control a desktop via API. GPT-5.2 is the best value OpenAI model for most teams.
GPT-5.5 costs $5/1M input and $30/1M output tokens. GPT-5.4 is $2.50/$15. GPT-5.2 is $1.50/$7.50. GPT-5.2 Mini is $0.15/$0.60 — the cheapest OpenAI option. ChatGPT Plus subscription is $20/month for consumer access.
It depends on the task. Claude Opus 4.7 (Anthropic) leads OpenAI on SWE-Bench Pro coding at 64.3% vs GPT-5.5's 58.6%. GPT-5.4 and GPT-5.5 are stronger for OpenAI-native agentic workflows and computer-use. For writing and general use, Claude Sonnet 4.6 and GPT-5.4 are closely matched.
No — all current OpenAI models (GPT-5.x, GPT-4o series) are closed-weight and available only through their API or ChatGPT. For open-weight alternatives, see Meta's Llama 4, Mistral, or DeepSeek models.
OpenAI's latest agentic flagship for coding, research, computer-use workflows, and long multi-step knowledge work.
OpenAI's latest flagship with unique desktop-control capabilities — it can see your screen, click, and navigate apps via the API.
GPT-5 Image is OpenAI's multimodal flagship optimized for deep visual understanding and generation tasks, built on the GPT-5 architecture with a 400K context window. It supersedes GPT-4o with significantly improved image reasoning, analysis, and generation capabilities.
GPT-5 Image Mini is OpenAI's mid-tier multimodal model optimized for image understanding and generation tasks at a balanced price point. It supersedes GPT-4o with improved visual reasoning capabilities while maintaining a large 400K context window.
OpenAI's most affordable production-grade model — faster and cheaper than GPT-4o with strong enough performance for the majority of everyday tasks.
GPT-5.1 is OpenAI's mid-tier flagship model, succeeding GPT-4o with improved reasoning, instruction-following, and a 400K context window at a competitive price point. It sits between GPT-4o and full GPT-5 in capability and cost.
Reliable OpenAI flagship for serious coding and product work — a strong default before GPT-5.4 was released.
GPT-4.1 is OpenAI's refined successor to GPT-4o, offering sharper instruction-following, stronger coding performance, and a massive 1M token context window at a mid-tier price point. It targets developers and power users who need reliable, precise outputs without paying flagship reasoning model prices.
Versatile multimodal model that handles image-related workflows and mixed-media prompts well.
Lower-cost OpenAI model that keeps a solid balance of usefulness, speed, and affordability for everyday tasks.
GPT-5 Mini is OpenAI's budget-tier distillation of GPT-5, designed for high-volume, cost-sensitive tasks that don't require full flagship reasoning depth. It supersedes GPT-4o with improved instruction following and a massively expanded 400K context window at a fraction of the cost.
GPT-4.1 Mini is OpenAI's cost-optimized small model from the GPT-4.1 family, designed to deliver strong instruction-following and coding performance at a fraction of flagship pricing. It targets high-volume, latency-sensitive applications where cost efficiency matters more than peak capability.
GPT-5 Chat is OpenAI's flagship conversational model, succeeding GPT-4o with improved reasoning, instruction-following, and multimodal capabilities. It targets professional and enterprise use cases where output quality matters more than cost.
o4 Mini is OpenAI's compact reasoning model that applies chain-of-thought thinking to complex problems at a fraction of the cost of o4. It delivers strong mathematical, coding, and logical reasoning capabilities while remaining accessible to developers on tighter budgets.
GPT-5 Pro is OpenAI's most capable flagship model, designed for complex reasoning, advanced coding, and high-stakes professional tasks. It supersedes GPT-4o with substantially improved intelligence at a premium price point reflecting its top-tier positioning.
GPT-5.1-Codex-Mini is OpenAI's budget-tier coding-specialized model built on the GPT-5.1 architecture, optimized for code generation, completion, and debugging at low cost. It offers a 400K context window, making it practical for large codebases without the price tag of flagship models.
OpenAI's o3 Mini is a compact reasoning model optimized for STEM tasks, offering chain-of-thought capabilities at a fraction of the cost of o3. It excels at math, coding, and logical problem-solving while maintaining a large 200K context window.
o4 Mini High is OpenAI's compact reasoning model running at its maximum reasoning effort setting, trading speed for deeper multi-step logical analysis. It applies extended chain-of-thought processing to complex problems while remaining significantly cheaper than full o3 or o4 class flagships.
OpenAI's o3 is a frontier reasoning model that uses extended chain-of-thought to solve complex problems in math, science, coding, and logic. It represents a significant step up from o1 in reasoning depth and accuracy.
GPT-5.1-Codex is OpenAI's coding-specialized flagship model, purpose-built for software development tasks with a massive 400K context window. It supersedes GPT-4o with deeper code comprehension, multi-file reasoning, and tighter integration with developer workflows.
o3 Mini High is OpenAI's compact reasoning model running at maximum reasoning effort, delivering deep chain-of-thought problem-solving in a cost-efficient package. It specializes in STEM tasks — math, coding, and logic — where extended deliberation yields significantly better results than standard chat models.
GPT-5 Codex is OpenAI's specialized coding-focused evolution of GPT-5, designed for software development tasks with a massive 400K context window for handling large codebases. It bridges the gap between raw language capability and developer-specific tooling, succeeding GPT-4o as OpenAI's primary coding workhorse.
GPT-4 Turbo is OpenAI's high-capability flagship model featuring a 128K context window, trained on data up to April 2024. It delivers strong reasoning, coding, and instruction-following across complex tasks.
GPT-5.1-Codex-Max is OpenAI's specialized coding-focused flagship model, built on the GPT-5 architecture with deep optimization for software development, code generation, and technical problem-solving. It supersedes GPT-4o with significantly improved code comprehension and a 400K context window suited for large codebases.
GPT-5.3 Chat is OpenAI's mid-cycle refinement of the GPT-5 series, offering improved instruction-following and reasoning over GPT-5.2 at a balanced price point. It targets professionals needing strong general-purpose performance without paying flagship model premiums.
GPT-5 Nano is OpenAI's smallest and fastest model in the GPT-5 family, optimized for high-throughput, low-latency tasks at near-minimal cost. It supersedes GPT-4o as the go-to option for lightweight inference at scale.
OpenAI's o3 Pro is the highest-tier reasoning model in the o3 family, designed for maximum accuracy on the most demanding intellectual tasks. It applies extended compute and deeper chain-of-thought reasoning to outperform standard o3 on math, science, coding, and complex analysis.
GPT-4.1 Nano is OpenAI's smallest and most cost-efficient model in the GPT-4.1 family, designed for high-throughput, latency-sensitive tasks at near-commodity pricing. It offers a 1M token context window at just $0.10/1M input tokens, making it one of the cheapest large-context models available.
o4 Mini Deep Research is OpenAI's cost-efficient reasoning model specialized for autonomous multi-step research tasks, capable of browsing the web, synthesizing sources, and producing detailed research reports. It brings deep research capabilities to a mid-tier price point by trading some of o4's raw power for significantly lower inference costs.
GPT-5.1 Chat is OpenAI's mid-tier conversational model, positioned as a capable successor to GPT-4o with improved instruction-following, reasoning, and knowledge depth at a balanced price point.
o1-pro is OpenAI's highest-tier reasoning model, running o1 with extended compute time for deeper, more reliable problem-solving on complex tasks. It is designed for users who need maximum accuracy and thoroughness over speed.
GPT-5.3-Codex is OpenAI's specialized coding-focused model in the GPT-5 lineage, built for deep software engineering tasks including code generation, debugging, and repository-level reasoning. It succeeds GPT-5.2 with improved instruction-following for complex multi-file codebases and a significantly expanded 400K context window.
GPT-4 Turbo Preview is an early access version of GPT-4 Turbo, OpenAI's then-flagship model featuring a 128K context window and knowledge improvements over the original GPT-4. It was designed to deliver GPT-4-class reasoning at reduced cost compared to the original GPT-4.
OpenAI's o3 Deep Research is a reasoning-heavy model purpose-built for multi-step research tasks, capable of autonomously browsing the web, synthesizing sources, and producing detailed analytical reports. It combines o3's chain-of-thought reasoning with agentic tool use to tackle complex, open-ended research questions.
GPT-4 Turbo (v1106) is an older snapshot of OpenAI's flagship GPT-4 Turbo model released in November 2023, offering a 128K context window with strong general-purpose reasoning and instruction-following capabilities. It predates later GPT-4 Turbo updates and GPT-4o, making it a legacy choice for workflows locked to this specific version.
OpenAI's o1 is a reasoning-focused model that uses chain-of-thought processing to tackle complex, multi-step problems in math, science, and coding. It deliberately 'thinks before answering,' trading speed for significantly improved accuracy on hard problems.
GPT Audio Mini is OpenAI's cost-efficient audio-capable model that handles real-time speech input and output alongside text, built on the GPT-4o Mini architecture. It's designed for voice-driven applications where low latency and affordable pricing matter more than peak intelligence.
GPT Audio is OpenAI's speech-capable model variant optimized for real-time audio input and output, enabling natural voice conversations and audio processing. It extends GPT-4o's multimodal capabilities with native audio understanding and generation without requiring separate transcription pipelines.
GPT-3.5 Turbo is OpenAI's legacy fast and affordable chat model, optimized for dialogue and straightforward text tasks at low cost. It was the backbone of early ChatGPT and remains a go-to for high-volume, cost-sensitive deployments.
A 20-billion parameter open-weights safety-focused model from OpenAI, designed primarily for content moderation, policy enforcement, and safeguard classification tasks. It is purpose-built to detect harmful, policy-violating, or unsafe content rather than serve as a general-purpose assistant.
GPT-4 is OpenAI's original flagship large language model, released in March 2023, offering strong reasoning and instruction-following across text tasks. It represents the foundational GPT-4 release before later variants like GPT-4 Turbo or GPT-4o improved speed, cost, and context length.
GPT-3.5 Turbo 16k is OpenAI's extended-context variant of their older flagship chat model, offering double the context window of the base 3.5 Turbo at a higher price point. It handles general-purpose text tasks but has been largely superseded by newer, more capable models.
An older versioned snapshot of GPT-3.5 Turbo (v0613), OpenAI's once-dominant mid-tier language model optimized for fast chat completions and instruction following. This specific checkpoint is frozen in time, predating later capability improvements introduced in subsequent GPT-3.5 Turbo updates.
GPT-3.5 Turbo Instruct is a legacy completion-style model from OpenAI, designed for instruction-following tasks using the older text completion API rather than the chat API. It excels at structured text generation, fill-in-the-middle tasks, and traditional NLP workflows that predate the chat paradigm.
GPT-4 v0314 is a frozen snapshot of the original GPT-4 release from March 2023, preserved for reproducibility and regression testing. It offers the same core reasoning capabilities as early GPT-4 but lacks all subsequent improvements, fine-tuning updates, and safety refinements.
| $2.50/1M |
| $2.00/1M |
| 400K |
| Fast |
| GPT-4o Mini Budget | $0.15/1M | $0.60/1M | 128K | Very fast |
| OpenAI: GPT-5.1 Balanced | $1.25/1M | $10.00/1M | 400K | Balanced |
| GPT-5.2 Premium | $1.75/1M | $14.00/1M | 200K | Balanced |
| OpenAI: GPT-4.1 Balanced | $2.00/1M | $8.00/1M | 1.047576M | Balanced |
| GPT-4o Balanced | $0.15/1M | $0.60/1M | 128K | Fast |
| GPT-5.2 Mini Balanced | $1.20/1M | $4.80/1M | 128K | Fast |
| OpenAI: GPT-5 Mini Budget | $0.25/1M | $2.00/1M | 400K | Very fast |
| OpenAI: GPT-4.1 Mini Budget | $0.40/1M | $1.60/1M | 1.047576M | Very fast |
| OpenAI: GPT-5 Chat Balanced | $1.25/1M | $10.00/1M | 128K | Balanced |
| OpenAI: o4 Mini Balanced | $1.10/1M | $4.40/1M | 200K | Deliberate |
| OpenAI: GPT-5 Pro Premium | $15.00/1M | $120.00/1M | 400K | Deliberate |
| OpenAI: GPT-5.1-Codex-Mini Budget | $0.25/1M | $2.00/1M | 400K | Very fast |
| OpenAI: o3 Mini Balanced | $1.10/1M | $4.40/1M | 200K | Deliberate |
| OpenAI: o4 Mini High Balanced | $1.10/1M | $4.40/1M | 200K | Deliberate |
| OpenAI: o3 Balanced | $2.00/1M | $8.00/1M | 200K | Deliberate |
| OpenAI: GPT-5.1-Codex Balanced | $1.25/1M | $10.00/1M | 400K | Balanced |
| OpenAI: o3 Mini High Balanced | $1.10/1M | $4.40/1M | 200K | Deliberate |
| OpenAI: GPT-5 Codex Balanced | $1.25/1M | $10.00/1M | 400K | Balanced |
| OpenAI: GPT-4 Turbo Premium | $10.00/1M | $30.00/1M | 128K | Balanced |
| OpenAI: GPT-5.1-Codex-Max Balanced | $1.25/1M | $10.00/1M | 400K | Balanced |
| OpenAI: GPT-5.3 Chat Balanced | $1.75/1M | $14.00/1M | 128K | Balanced |
| OpenAI: GPT-5 Nano Budget | $0.05/1M | $0.40/1M | 400K | Very fast |
| OpenAI: o3 Pro Premium | $20.00/1M | $80.00/1M | 200K | Deliberate |
| OpenAI: GPT-4.1 Nano Budget | $0.10/1M | $0.40/1M | 1.047576M | Very fast |
| OpenAI: o4 Mini Deep Research Balanced | $2.00/1M | $8.00/1M | 200K | Deliberate |
| OpenAI: GPT-5.1 Chat Balanced | $1.25/1M | $10.00/1M | 128K | Balanced |
| OpenAI: o1-pro Premium | $150.00/1M | $600.00/1M | 200K | Deliberate |
| OpenAI: GPT-5.3-Codex Balanced | $1.75/1M | $14.00/1M | 400K | Balanced |
| OpenAI: GPT-4 Turbo Preview Premium | $10.00/1M | $30.00/1M | 128K | Balanced |
| OpenAI: o3 Deep Research Premium | $10.00/1M | $40.00/1M | 200K | Deliberate |
| OpenAI: GPT-4 Turbo (older v1106) Premium | $10.00/1M | $30.00/1M | 128K | Balanced |
| OpenAI: o1 Premium | $15.00/1M | $60.00/1M | 200K | Deliberate |
| OpenAI: GPT Audio Mini Balanced | $0.60/1M | $2.40/1M | 128K | Fast |
| OpenAI: GPT Audio Balanced | $2.50/1M | $10.00/1M | 128K | Balanced |
| OpenAI: GPT-3.5 Turbo Budget | $0.50/1M | $1.50/1M | 16K | Very fast |
| OpenAI: gpt-oss-safeguard-20b Budget | $0.07/1M | $0.30/1M | 131K | Fast |
| OpenAI: GPT-4 Premium | $30.00/1M | $60.00/1M | 8K | Balanced |
| OpenAI: GPT-3.5 Turbo 16k Balanced | $3.00/1M | $4.00/1M | 16K | Fast |
| OpenAI: GPT-3.5 Turbo (older v0613) Balanced | $1.00/1M | $2.00/1M | 4K | Very fast |
| OpenAI: GPT-3.5 Turbo Instruct Balanced | $1.50/1M | $2.00/1M | 4K | Very fast |
| OpenAI: GPT-4 (older v0314) Premium | $30.00/1M | $60.00/1M | 8K | Balanced |
