ChatGPT Models: Full List, Key Features & Use Cases

OpenAI's ChatGPT model family has grown from a single research preview into a sprawling lineup spanning multiple generations, reasoning tiers, and image models. This guide lists every major ChatGPT model in order from the newest to the oldest, with a concise overview of each model's specs, key features, and intended use — plus links to try the ones available on Chat Smith.

Released September 3, 2026 and generally available the following day, GPT-6 Astra is OpenAI's current frontier model and the largest capability jump in the lineup since GPT-5. It is state of the art on computer use, browsing, software engineering, cybersecurity, science and professional work, scoring 98% on FrontierMath Tier 4 and 99.9% on ARC-AGI-3. Context window: 1.05M tokens, 128K max output. Pricing: $10/1M input, $50/1M output. It is also the first OpenAI model to reach the Critical cybersecurity threshold under the company's Preparedness Framework, so the public release ships with extra safeguards on sensitive requests. Best for: long-horizon agentic work, computer and browser use, and end-to-end professional tasks.

Released July 9, 2026 after a limited preview, GPT-5.6 Sol is the flagship of the GPT-5.6 family. The release introduced a new naming system in which the number marks the generation while Sol, Terra and Luna mark durable capability tiers that can advance on their own cadence. Sol is built for complex reasoning, large codebases and long-running agentic work. Context window: 1.05M tokens, 128K max output. Pricing: $5/1M input, $30/1M output. Best for: the hardest reasoning and coding tasks, where peak capability matters more than cost.

Terra is the balanced middle tier of the GPT-5.6 family, released alongside Sol and Luna on July 9, 2026. OpenAI positions it as competitive with GPT-5.5 at a fraction of the cost, which makes it the sensible default for everyday interactive and agentic work. It shares the family's 1.05M-token context window and 128K max output. OpenAI cut Terra's price 20% on July 30, 2026, to $2/1M input and $12/1M output. Best for: everyday production traffic where Sol would be more than the task needs.

GPT-5.6 Luna is the fastest and lowest-cost tier in the family, built for high-volume, latency-sensitive work. It shares the same 1.05M-token context window as Sol and Terra, so choosing a tier is a cost and quality decision rather than a context-length one. OpenAI cut Luna's price 80% on July 30, 2026, to $0.20/1M input and $1.20/1M output, putting it at roughly one twenty-fifth of Sol's rate. Best for: classification, summarisation, subagents and high-volume automation loops.

Released April 23, 2026, GPT-5.5 was the first OpenAI model to ship a 1M-token API context window, matching GPT-5.4's per-token latency while using fewer tokens per task. It scored 82.7% on Terminal-Bench 2.0 and 35.4% on FrontierMath Tier 4. Standard pricing: $5/1M input, $30/1M output; GPT-5.5 Pro runs $30/$180. Best for: agentic coding, research, document creation.

Launched April 21–22, 2026, GPT Image 2 is OpenAI's third-generation native image model and the first to reason before generating: it plans and self-checks a prompt before rendering pixels. It outputs up to 2K resolution, up to eight images per prompt, and renders multilingual text far more reliably than earlier versions. Best for: high-fidelity image generation, brand visuals, precise multi-turn edits.

Released March 5, 2026 in Thinking and Pro variants, with mini and nano variants following on March 17, GPT-5.4 was praised for a low tendency to hallucinate but criticized for occasionally not following prompts precisely. Standard pricing: $2.50/1M input, $15.00/1M output. Best for: general professional work, everyday coding and writing before the 5.5 upgrade.

Released December 16, 2025 and rolled out globally as "ChatGPT Images," GPT Image 1.5 refined the original GPT Image 1 with more precise edits and better prompt fidelity, while keeping the model's core strength: layout and instruction-following that outperformed DALL-E 3. Best for: quick image edits, everyday creative generation.

Released December 11, 2025 amid intense competition with Google's Gemini 3 Pro, GPT-5.2 shipped in Instant, Thinking, and Pro modes plus a dedicated GPT-5.2-Codex variant. It was the first OpenAI model to cross 90% on ARC-AGI-1. Context window: 400K tokens, 128K max output. Pricing: $1.75/1M input, $14.00/1M output. Best for: complex reasoning and coding at a lower cost than GPT-5.4/5.5.

Released as a family of five models starting November 12, 2025, GPT-5.1's headline feature was personalization: eight selectable personality options and a warmer default tone compared to GPT-5. It was primarily a refinement release rather than a capability leap. Best for: everyday ChatGPT conversations where tone and personality matter.

Released August 7, 2025, GPT-5 unified OpenAI's chat and reasoning lines into a single system with a built-in router, replacing the separate GPT-4o and o-series lineup. Context window: 400K tokens. Pricing: $0.625/1M input, $5.00/1M output. It also shipped in GPT-5 mini and GPT-5 Nano sizes for cheaper, faster workloads. Best for: general-purpose chat, coding, and reasoning in one model.

Introduced to the API on April 23, 2025, GPT Image 1 was OpenAI's first native image model built directly into GPT rather than bolted on as a separate tool like DALL-E 3. It drove such heavy demand at launch that Sam Altman said the company's GPUs were "melting." A cheaper GPT Image 1 mini variant followed in October 2025, priced roughly 80% below the original. Best for: legacy image generation workflows, now succeeded by GPT Image 1.5 and 2.

Released April 16, 2025, GPT o3 was OpenAI's reasoning-focused successor to o1, with a 200K-token context window and strong results on STEM, coding, and visual reasoning tasks. Initial API pricing of $10/1M input and $40/1M output was cut to $2/$8 two months later. Best for: multi-step reasoning, math, and visual analysis where accuracy matters more than speed.

Released April 14, 2025, GPT-4.1 replaced GPT-4o as OpenAI's recommended default model. It was cheaper ($2.00/1M input, $8.00/1M output vs. GPT-4o's $2.50/$10.00), added a 1M-token context window versus GPT-4o's 128K, and scored better on instruction-following and coding benchmarks. Best for: general-purpose tasks needing long context on a budget.

Released February 27, 2025 under the codename "Orion," GPT-4.5 was OpenAI's last flagship chat model without chain-of-thought reasoning. Reviewers saw it as an incremental step over GPT-4o rather than a major leap, and OpenAI phased it out of the API on July 14, 2025 in favor of GPT-4.1. Best for: reference only — superseded by GPT-4.1 and the GPT-5 line.

Released May 13, 2024, GPT-4o ("o" for omni) was OpenAI's first natively multimodal flagship, processing and generating text, image, and audio in one model. Context window: 128K tokens. Pricing: $2.50/1M input, $10.00/1M output. A cheaper GPT-4o mini variant served high-volume workloads. It was removed from ChatGPT in February 2026 over sycophancy concerns. Best for: legacy multimodal workflows still on GPT-4o.

Released March 14, 2023, GPT-4 powered the original ChatGPT Plus launch and later gained image understanding through GPT-4V. With an 8K token context window (extended to 32K in a later variant), it was the model that convinced much of the world GPT-class systems were ready for serious professional use. Best for: historical reference — fully superseded by GPT-4o and later models.

Released November 30, 2022 as the model behind the original ChatGPT launch, GPT-3.5 is the release that turned large language models into a mainstream consumer product almost overnight. It's long since been superseded on every axis of capability, but it remains the historical starting point for the entire ChatGPT product line. Best for: historical reference only.

OpenAI's ChatGPT lineup now spans five major generations and multiple capability tiers, from budget-friendly nano and Luna models to frontier reasoning with Sol and Astra. For teams and individuals who want to use these models without managing separate API keys, Chat Smith gives you access to the available GPT models alongside Claude Sonnet 5, Gemini 3.5 Flash, Grok 4.5 and DeepSeek V4 Pro — all from a single interface. Browse the full model catalog on Chat Smith to find the right model for your workflow.