{"id":6418,"date":"2025-12-12T00:14:24","date_gmt":"2025-12-12T04:14:24","guid":{"rendered":"https:\/\/wp.glbgpt.com\/?p=6418"},"modified":"2026-10-06T04:04:45","modified_gmt":"2026-10-06T08:04:45","slug":"chatgpt-5-2","status":"publish","type":"post","link":"https:\/\/wp.glbgpt.com\/hub\/chatgpt-5-2","title":{"rendered":"ChatGPT\u202f5.2: 3\u00d7 Smarter, Perfect Scores Across the Board &#8211; The King Is Back"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">GPT-5.2 is an OpenAI model for professional work. In ChatGPT, its Instant, Thinking and Pro variants retired on June 12, 2026. The base <code>gpt-5.2<\/code> API documentation remains available, and GlobalGPT has a dedicated GPT-5.2 workspace. The right path depends on whether you need ChatGPT, an API identifier or a multi-model workspace.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">If your work moves from research to writing and coding, <a href=\"https:\/\/www.glbgpt.com\/home?inviter=hub_content_home&amp;login=1\">GlobalGPT<\/a> brings multiple AI models into one subscription. Use research notes to shape a draft, then move into code or supporting visuals as the project needs. The value is having those stages available in one workspace.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img alt=\"\" fetchpriority=\"high\" decoding=\"async\" width=\"1024\" height=\"547\" src=\"https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-1024x547.png\" alt=\"\" class=\"wp-image-20305\" srcset=\"https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-1024x547.png 1024w, https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-300x160.png 300w, https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-768x410.png 768w, https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-1536x820.png 1536w, https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-18x10.png 18w, https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2026\/10\/image-2048x1094.png 2048w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<div class=\"wp-block-buttons is-content-justification-center is-layout-flex wp-container-core-buttons-is-layout-3e41869c wp-block-buttons-is-layout-flex\">\n<div class=\"wp-block-button is-style-outline is-style-outline--1\"><a class=\"wp-block-button__link has-blush-light-purple-gradient-background has-background wp-element-button\" href=\"https:\/\/www.glbgpt.com\/home\/gpt-6-1-sol\"><strong>Try GPT-6.1 Sol on GlobalGPT<\/strong><\/a><\/div>\n<\/div>\n\n\n\n<nav aria-label=\"Table of contents\" style=\"box-sizing:border-box;width:100%;max-width:860px;margin:24px auto;padding:20px 24px;border:1px solid #c5d9e0;border-radius:14px;background:#eef6f8;color:#173e4e\"><p style=\"font-size:19px;font-weight:700;margin:0 0 12px\">On this page<\/p><ul style=\"padding-left:22px;margin:0;line-height:1.85\"><li><a href=\"#quick-answer\" style=\"color:#145d78\">GPT-5.2 at a glance<\/a><\/li><li><a href=\"#what-is-gpt-5-2\" style=\"color:#145d78\">What is GPT-5.2?<\/a><\/li><li><a href=\"#current-access\" style=\"color:#145d78\">Where can you use GPT-5.2?<\/a><\/li><li><a href=\"#model-versions\" style=\"color:#145d78\">Instant, Thinking and Pro: what the names meant<\/a><\/li><li><a href=\"#launch-benchmarks\" style=\"color:#145d78\">What did GPT-5.2 improve at launch?<\/a><\/li><li><a href=\"#practical-use\" style=\"color:#145d78\">What the scores mean for real work<\/a><\/li><li><a href=\"#api-model-names\" style=\"color:#145d78\">API model names and pricing<\/a><\/li><li><a href=\"#faq\" style=\"color:#145d78\">Frequently asked questions<\/a><\/li><li><a href=\"#next-step\" style=\"color:#145d78\">Choose the right access path<\/a><\/li><\/ul><\/nav>\n\n\n\n<h2 id=\"quick-answer\" class=\"wp-block-heading\">GPT-5.2 at a glance<\/h2>\n\n\n\n<section aria-label=\"GPT-5.2 at a glance\" style=\"box-sizing:border-box;width:100%;max-width:860px;margin:24px auto;border:1px solid #c5d9e0;border-left:5px solid #215b72;border-radius:14px;background:#fff;color:#233f4b;overflow:hidden\"><div style=\"overflow-x:auto\"><table style=\"width:100%;border-collapse:collapse;line-height:1.6;font-size:16px\"><caption style=\"text-align:left;font-weight:700;padding:16px;color:#163e4c\">GPT-5.2 at a glance<\/caption><thead><tr><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">Question<\/th><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">Short answer<\/th><\/tr><\/thead><tbody><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">What is GPT-5.2?<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">An OpenAI model whose base API documentation now labels it a previous flagship for professional work.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Can you still use it in ChatGPT?<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">No. OpenAI\u2019s June 12, 2026 release notes retire GPT-5.2 Instant, Thinking and Pro from ChatGPT.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Is the API the same as ChatGPT?<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">No. The base gpt-5.2 API page, the retired chat-latest alias, and ChatGPT access are separate records.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">What does the API cost?<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">$1.75 input, $0.175 cached input and $14 output per 1,000,000 text tokens on the standard price view checked here.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Does one benchmark prove general intelligence?<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">No. The launch results are dated, vendor-reported evaluations with different tools, models and research conditions.<\/td><\/tr><\/tbody><\/table><\/div><p style=\"padding:0 16px 16px;margin:14px 0 0;font-size:14px;color:#45616b\">Sources: <a href=\"https:\/\/help.openai.com\/en\/articles\/6825453-chatgpt-release-notes\" style=\"color:#215b72\">OpenAI release notes<\/a> and <a href=\"https:\/\/developers.openai.com\/api\/docs\/models\/gpt-5.2\" style=\"color:#215b72\">GPT-5.2 model documentation<\/a>. API rates are not ChatGPT or GlobalGPT subscription prices.<\/p><\/section>\n\n\n\n<h2 id=\"what-is-gpt-5-2\" class=\"wp-block-heading\">What is GPT-5.2?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">GPT-5.2 is an OpenAI model that its <a href=\"https:\/\/developers.openai.com\/api\/docs\/models\/gpt-5.2\">API documentation<\/a> now describes as a previous flagship for professional work. The documented base model accepts text and image inputs and produces text. GPT-5.2 is a model name; ChatGPT is a product in which model access can change over time.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That distinction matters when you follow an older guide. A model\u2019s API specifications do not establish which ChatGPT plan includes it, and a retired API identifier does not describe every product that once used the GPT-5.2 name.<\/p>\n\n\n\n<h2 id=\"current-access\" class=\"wp-block-heading\">Can you still use GPT-5.2 in ChatGPT?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>GPT-5.2 Instant, GPT-5.2 Thinking and GPT-5.2 Pro are no longer available in ChatGPT.<\/strong> OpenAI\u2019s <a href=\"https:\/\/help.openai.com\/en\/articles\/6825453-chatgpt-release-notes\">June 12, 2026 release notes<\/a> state that retirement date explicitly. The notice said existing conversations would continue on the corresponding GPT-5.5 model; that describes the migration announced in June, not which model is newest today.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/static.futureshareai.com\/glb_features\/mcp\/3\/chatgpt-52-chat-retirement_9101571143064e8384e68663ff62edcb.webp\" alt=\"OpenAI June 12, 2026 notice retiring GPT-5.2 Instant, Thinking and Pro from ChatGPT.\"\/><figcaption class=\"wp-element-caption\">OpenAI\u2019s June 12, 2026 ChatGPT retirement notice. Source: <a href=\"https:\/\/help.openai.com\/en\/articles\/6825453-chatgpt-release-notes\">ChatGPT release notes<\/a>; captured September 28, 2026.<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">The notice concerns ChatGPT access. API model identifiers and token billing are documented separately, so the ChatGPT retirement date does not establish the status of every API model.<\/p>\n\n\n\n<aside id=\"access-status-summary\" style=\"box-sizing:border-box;margin:24px 0;padding:22px;border:1px solid #b9d2db;border-left:5px solid #256477;border-radius:12px;background:#f1f7f9;color:#183a48;line-height:1.65;\">\n<p style=\"margin:0 0 12px;color:#183a48;font-size:20px;font-weight:700;\">GPT-5.2 access depends on the product<\/p>\n<p style=\"margin:0 0 12px;color:#183a48;\"><strong>ChatGPT:<\/strong> OpenAI says GPT-5.2 Instant, Thinking and Pro are no longer available there as of June 12, 2026.<\/p>\n<p style=\"margin:0 0 12px;color:#183a48;\"><strong>GlobalGPT:<\/strong> open the dedicated GPT-5.2 workspace, or choose the GPT-6 Astra workspace for a newer GPT option.<\/p>\n<p style=\"margin:0;color:#405c68;font-size:14px;\">The ChatGPT notice does not establish availability in the API, Codex or another platform. <a href=\"https:\/\/help.openai.com\/en\/articles\/6825453-chatgpt-release-notes\" style=\"color:#145d78;text-decoration:underline;\">Read OpenAI\u2019s release notes<\/a>.<\/p>\n<\/aside>\n\n\n\n<h3 id=\"globalgpt-access\" class=\"wp-block-heading\">The GPT-5.2 page in GlobalGPT<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Open the <a href=\"https:\/\/www.glbgpt.com\/home\/gpt-5-2?inviter=hub_content_gpt52&amp;login=1\">GPT-5.2 workspace in GlobalGPT<\/a> to reach its dedicated model page. As of September 28, 2026, the page includes an \u201cAsk anything\u201d field, file-attachment and voice-input controls, and a Send button.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">GlobalGPT brings multiple models into one subscription for work that moves from research to drafting, coding and supporting visuals. For a newer GPT option, open the <a href=\"https:\/\/www.glbgpt.com\/home\/gpt-6-astra?inviter=hub_content_gpt6astra&amp;login=1\">GPT-6 Astra page<\/a>. Choose a model entry point around the task at hand, whether you are organizing research, shaping a draft or working on code.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/static.futureshareai.com\/glb_features\/mcp\/3\/globalgpt-gpt-52-workspace_6f1a6a05efe9498d832166ff8172f878.webp\" alt=\"GlobalGPT GPT-5.2 workspace showing the model name and Ask anything input controls.\"\/><figcaption class=\"wp-element-caption\">GlobalGPT\u2019s GPT-5.2 workspace, captured September 28, 2026, with the model label, input field and attachment controls.<\/figcaption><\/figure>\n\n\n\n<h2 id=\"model-versions\" class=\"wp-block-heading\">Instant, Thinking and Pro: what the names meant<\/h2>\n\n\n\n<section aria-label=\"Historical GPT-5.2 names and product boundaries\" style=\"box-sizing:border-box;width:100%;max-width:860px;margin:24px auto;border:1px solid #c5d9e0;border-left:5px solid #215b72;border-radius:14px;background:#fff;color:#233f4b;overflow:hidden\"><div style=\"overflow-x:auto\"><table style=\"width:100%;border-collapse:collapse;line-height:1.6;font-size:16px\"><caption style=\"text-align:left;font-weight:700;padding:16px;color:#163e4c\">Historical GPT-5.2 names and product boundaries<\/caption><thead><tr><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">Name<\/th><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">What it described<\/th><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">Boundary<\/th><\/tr><\/thead><tbody><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">GPT-5.2 Instant<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">A faster ChatGPT variant named in the 2025 launch.<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Retired from ChatGPT on June 12, 2026; the name does not define current API access.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">GPT-5.2 Thinking<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">A reasoning-oriented ChatGPT variant named in the launch and benchmark tables.<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Historical ChatGPT product name; launch benchmark conditions are separate from current account rights.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">GPT-5.2 Pro<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">The higher-effort variant used in selected launch evaluations.<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Do not confuse the model name with a ChatGPT Pro subscription.<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">gpt-5.2<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">The base model on the current API documentation page.<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">Documentation and snapshots do not prove that a request will succeed for a particular account.<\/td><\/tr><\/tbody><\/table><\/div><p style=\"padding:0 16px 16px;margin:14px 0 0;font-size:14px;color:#45616b\">This is a historical naming table. It does not claim that the three ChatGPT variants are currently selectable.<\/p><\/section>\n\n\n\n<h2 id=\"launch-benchmarks\" class=\"wp-block-heading\">What did GPT-5.2 improve at launch?<\/h2>\n\n\n\n<h2 id=\"historical-benchmark-method\" class=\"wp-block-heading\">How to read GPT-5.2\u2019s launch benchmarks<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The tables below preserve results from OpenAI\u2019s <a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-2\/\">December 11, 2025 GPT-5.2 announcement<\/a>. They are historical, vendor-reported evaluations. They do not rank today\u2019s models or describe new tests performed for this article.<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Reasoning effort:<\/strong> the appendix says GPT-5.2 Thinking and Pro ran at <code>xhigh<\/code>, while GPT-5.1 Thinking ran at <code>high<\/code>. The professional evaluations were an exception: GPT-5.2 Thinking used <code>heavy<\/code>. Keep any row-specific qualification with its result.<\/li>\n\n\n\n<li><strong>Tools:<\/strong> \u201cno tools,\u201d \u201cw\/ Python\u201d and \u201cw\/ search, Python\u201d are different test conditions. Compare matching conditions before interpreting a change in score.<\/li>\n\n\n\n<li><strong>Environment:<\/strong> OpenAI says these benchmarks ran in a research environment whose output may differ from production ChatGPT.<\/li>\n<\/ul>\n\n\n\n<section aria-label=\"OpenAI launch results that can be traced to the December 11, 2025 report\" style=\"box-sizing:border-box;width:100%;max-width:860px;margin:24px auto;border:1px solid #c5d9e0;border-left:5px solid #6d5b29;border-radius:14px;background:#fff;color:#233f4b;overflow:hidden\"><div style=\"overflow-x:auto\"><table style=\"width:100%;border-collapse:collapse;line-height:1.6;font-size:16px\"><caption style=\"text-align:left;font-weight:700;padding:16px;color:#163e4c\">OpenAI launch results that can be traced to the December 11, 2025 report<\/caption><thead><tr><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">Evaluation and condition<\/th><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">GPT-5.2 Thinking<\/th><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">GPT-5.2 Pro<\/th><th scope=\"col\" style=\"padding:12px;text-align:left;border-bottom:2px solid #b9d1db;background:#e8f2f5;color:#163e4c\">GPT-5.1 \/ GPT-5 comparison<\/th><\/tr><\/thead><tbody><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">GDPval \u2014 wins or ties, ties allowed<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">70.9%<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">74.1%<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">38.8% (GPT-5)<\/td><\/tr><tr><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">ARC-AGI-2 \u2014 Verified<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">52.9%<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">54.2% (high)<\/td><td style=\"padding:12px;border-bottom:1px solid #d7e3e8;vertical-align:top\">17.6% (GPT-5.1 Thinking)<\/td><\/tr><\/tbody><\/table><\/div><p style=\"padding:0 16px 16px;margin:14px 0 0;font-size:14px;color:#45616b\">OpenAI reported these figures in its December 11, 2025 announcement and appendix. They are vendor-reported historical results, not an independent reproduction or a cross-vendor comparison. The GDPval row measures wins or ties against industry professionals; the ARC Pro row keeps its high-effort condition visible. See the <a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-2\/\" style=\"color:#215b72\">original report<\/a>.<\/p><\/section>\n\n\n\n<h3 id=\"historical-coding\" class=\"wp-block-heading\">Coding: separate tests answer different questions<\/h3>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2025\/12\/image-39-1024x263.png\" alt=\"OpenAI historical coding benchmarks for GPT-5.2 Thinking, GPT-5.2 Pro and GPT-5.1 Thinking.\"\/><figcaption class=\"wp-element-caption\">Coding results in the <a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-2\/\">December 11, 2025 OpenAI appendix<\/a>. The original table is retained; the SWE-Lancer asterisk is explained below.<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">GPT-5.2 Thinking scored 55.6% on SWE-Bench Pro, Public and 80.0% on SWE-bench Verified, compared with 50.8% and 76.3% for GPT-5.1 Thinking. Those are two different benchmarks, so the scores should not be combined into a single coding success rate.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For SWE-Lancer, IC Diamond, the table reports 74.6% versus 69.7%. OpenAI says it omitted <strong>40 of 237 problems<\/strong> that did not run on its infrastructure. That exclusion is part of the result\u2019s scope; the figure is not a score over all 237 problems.<\/p>\n\n\n\n<h3 id=\"historical-academic\" class=\"wp-block-heading\">Academic results: a perfect score on one test<\/h3>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2025\/12\/image-40-1024x490.png\" alt=\"OpenAI historical academic benchmark table, including AIME 2025 without tools and FrontierMath with Python.\"\/><figcaption class=\"wp-element-caption\">Academic results from <a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-2\/\">OpenAI\u2019s December 11, 2025 report<\/a>, with the original AIME highlight retained. Tool conditions remain visible in each row.<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">GPT-5.2 Thinking and Pro each reached 100.0% on <strong>AIME 2025 (no tools)<\/strong>. That result applies to this test. The same table reports 34.5% and 36.6% on HLE without tools, and 45.5% and 50.0% with search and Python. These results show why AIME alone cannot establish accuracy across every task.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The FrontierMath rows also need their full labels: GPT-5.2 Thinking scored 40.3% on Tier 1\u20133 and 14.6% on Tier 4, both with Python. A reader evaluating mathematical work should keep the problem set and tool access attached to each number.<\/p>\n\n\n\n<h3 id=\"historical-vision\" class=\"wp-block-heading\">Vision: keep the Python conditions visible<\/h3>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/wp.glbgpt.com\/wp-content\/uploads\/2025\/12\/image-41-1024x381.png\" alt=\"OpenAI historical vision results, including CharXiv with and without tools and ScreenSpot Pro with Python.\"\/><figcaption class=\"wp-element-caption\">Vision results from <a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-2\/\">OpenAI\u2019s December 11, 2025 appendix<\/a>. \u201cNo tools\u201d and \u201cw\/ Python\u201d remain separate conditions; dashes are unreported values, not zero scores.<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">For GPT-5.2 Thinking, <strong>CharXiv reasoning (no tools)<\/strong> is 82.1%; <strong>CharXiv reasoning (w\/ Python)<\/strong> is 88.7%. The corresponding GPT-5.1 Thinking results are 67.0% and 80.3%. Removing the tool labels would make these comparisons ambiguous.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>ScreenSpot Pro (w\/ Python)<\/strong> is 86.3% for GPT-5.2 Thinking versus 64.2% for GPT-5.1 Thinking. The table also separates MMMU Pro without tools, MMMU Pro with Python and Video MMMU without tools. These are vision-understanding evaluations, not evidence of native video generation.<\/p>\n\n\n\n<h2 id=\"practical-use\" class=\"wp-block-heading\">What a coding case can tell you<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">A useful historical example comes from <a href=\"https:\/\/simonwillison.net\/2025\/Dec\/15\/porting-justhtml\/\">Simon Willison\u2019s December 15, 2025 account of porting JustHTML from Python to JavaScript<\/a>. He used GPT-5.2 through Codex CLI, with an existing implementation and the html5lib test suite to check the port. He also described a pause after reaching his usage allowance.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The practical lesson is to give a coding task a checkable target: expected behavior, representative inputs and tests that can catch regressions. This was one author\u2019s project in a specific tool environment. It does not establish a general success rate or describe a test performed in GlobalGPT.<\/p>\n\n\n\n<h2 id=\"api-model-names\" class=\"wp-block-heading\">Check the exact API model name<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">For an existing integration, check the full model identifier before applying an old setup guide. The OpenAI documentation distinguishes these names:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><code>gpt-5.2<\/code>: the base model documented for professional work.<\/li>\n\n\n\n<li><code>gpt-5.2-2025-12-11<\/code>: the snapshot listed on that model\u2019s documentation page.<\/li>\n\n\n\n<li><code>gpt-5.2-chat-latest<\/code>: a separate identifier with an August 10, 2026 shutdown date in the <a href=\"https:\/\/developers.openai.com\/api\/docs\/deprecations\">API deprecation record<\/a>.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">That shutdown entry concerns <code>gpt-5.2-chat-latest<\/code>. The base model\u2019s documentation remains a source for its specifications, but a documentation page alone does not confirm that a request will succeed for your account. Check the model identifier and your account\u2019s access before using it in a workflow.<\/p>\n\n\n\n<h2 id=\"api-pricing\" class=\"wp-block-heading\">GPT-5.2 API pricing and limits<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The base GPT-5.2 API has separate rates for uncached input, cached input and output. Use the rate for each category when estimating a request; do not apply the cheapest category to every token.<\/p>\n\n\n\n<figure id=\"d01-api-rates\" aria-labelledby=\"d01-title\" aria-describedby=\"d01-scope\" style=\"box-sizing:border-box;max-width:100%;margin:28px 0;padding:clamp(18px,4vw,30px);border:1px solid #cadce2;border-radius:16px;background:#f3f8fa;color:#173e4e;font-family:Arial,Helvetica,sans-serif;font-size:16px;line-height:1.6;\">\n<figcaption id=\"d01-title\" style=\"margin:0 0 7px;color:#173e4e;font-size:22px;line-height:1.3;font-weight:700;\">GPT-5.2 API: the cost of each token type<\/figcaption>\n<p id=\"d01-scope\" style=\"margin:0 0 24px;color:#405d69;font-size:14px;\">Standard text-token rates in USD per 1,000,000 tokens. Checked September 30, 2026.<\/p>\n<div style=\"margin:0 0 20px;\">\n<div style=\"display:flex;flex-wrap:wrap;justify-content:space-between;gap:4px 16px;margin:0 0 7px;color:#173e4e;\"><span>Uncached input<\/span><strong style=\"color:#173e4e;font-variant-numeric:tabular-nums;\">$1.75<\/strong><\/div>\n<div aria-hidden=\"true\" style=\"width:100%;height:16px;background:#dce7eb;border-left:1px solid #8398a1;box-sizing:border-box;\"><div style=\"width:12.500%;height:100%;background:#326b83;\"><\/div><\/div>\n<\/div>\n<div style=\"margin:0 0 20px;\">\n<div style=\"display:flex;flex-wrap:wrap;justify-content:space-between;gap:4px 16px;margin:0 0 7px;color:#173e4e;\"><span>Cached input<\/span><strong style=\"color:#173e4e;font-variant-numeric:tabular-nums;\">$0.175<\/strong><\/div>\n<div aria-hidden=\"true\" style=\"width:100%;height:16px;background:#dce7eb;border-left:1px solid #8398a1;box-sizing:border-box;\"><div style=\"width:1.2500%;height:100%;background:#487c6d;\"><\/div><\/div>\n<\/div>\n<div style=\"margin:0 0 20px;\">\n<div style=\"display:flex;flex-wrap:wrap;justify-content:space-between;gap:4px 16px;margin:0 0 7px;color:#173e4e;\"><span>Output<\/span><strong style=\"color:#173e4e;font-variant-numeric:tabular-nums;\">$14.00<\/strong><\/div>\n<div aria-hidden=\"true\" style=\"width:100%;height:16px;background:#dce7eb;border-left:1px solid #8398a1;box-sizing:border-box;\"><div style=\"width:100%;height:100%;background:#173e4e;\"><\/div><\/div>\n<\/div>\n<div aria-hidden=\"true\" style=\"display:flex;justify-content:space-between;gap:8px;margin:0 0 20px;color:#405d69;font-size:13px;font-variant-numeric:tabular-nums;\"><span>$0<\/span><span>$7<\/span><span>$14<\/span><\/div>\n<p style=\"margin:0 0 10px;padding-top:16px;border-top:1px solid #cadce2;color:#173e4e;\"><strong style=\"color:#173e4e;\">Estimate output separately.<\/strong> Every bar starts at zero on the same $0\u2013$14 scale. Input and output have different rates, so total token count alone is not enough to estimate cost.<\/p>\n<p style=\"margin:0;color:#405d69;font-size:13px;\">Source: <a href=\"https:\/\/developers.openai.com\/api\/docs\/models\/gpt-5.2\" style=\"color:#145d78;text-decoration:underline;\">OpenAI GPT-5.2 model documentation<\/a>. These are API usage rates; ChatGPT subscriptions and GlobalGPT subscriptions use separate billing. Batch discounts and additional tool fees are outside this chart.<\/p>\n<\/figure>\n\n\n\n<h3 id=\"api-cost-example\" class=\"wp-block-heading\">A simple cost example<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">For a hypothetical workload with <strong>1,000,000 uncached input tokens<\/strong> and <strong>100,000 output tokens<\/strong>, the calculation is $1.75 + $1.40 = <strong>$3.15<\/strong>. This is an arithmetic example, not a measured bill. It excludes additional tool fees and assumes none of the input is billed at the cached-input rate.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Long answers can add noticeably to the total because output tokens have a higher unit rate. Set the requested answer length around the task, then use the recorded input, cached-input and output usage to check the cost.<\/p>\n\n\n\n<h3 id=\"api-limits\" class=\"wp-block-heading\">Documented API limits<\/h3>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Context window:<\/strong> 400,000 tokens.<\/li>\n\n\n\n<li><strong>Maximum output:<\/strong> 128,000 tokens.<\/li>\n\n\n\n<li><strong>Input and output:<\/strong> text and image inputs; text output.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">These specifications apply to the base <code>gpt-5.2<\/code> API model. The <a href=\"https:\/\/developers.openai.com\/api\/docs\/models\/gpt-5.2\">same documentation<\/a> lists separate rate limits by API usage tier. Model capacity, request-rate limits and ChatGPT message allowances are different limits.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Use GlobalGPT when a project moves from research into writing, coding or supporting visuals and you prefer one workspace for several models. The GlobalGPT CLI connects these AI workflows with your terminal, development tools and existing production tools. Start with the model page that matches the task, then switch when the next step calls for it.<\/p>\n\n\n\n<h2 id=\"faq\" class=\"wp-block-heading\">Frequently asked questions<\/h2>\n\n\n\n<h3 id=\"is-gpt-5-2-still-available-in-chatgpt\" class=\"wp-block-heading\">Is GPT-5.2 still available in ChatGPT?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>No.<\/strong> OpenAI\u2019s <a href=\"https:\/\/help.openai.com\/en\/articles\/6825453-chatgpt-release-notes\">June 12, 2026 release notes<\/a> say GPT-5.2 Instant, Thinking and Pro are no longer available in ChatGPT. That statement is about ChatGPT, not every API identifier or another platform.<\/p>\n\n\n\n<h3 id=\"can-i-still-use-gpt-5-2-through-the-api\" class=\"wp-block-heading\">Can I still use GPT-5.2 through the API?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">The base <code>gpt-5.2<\/code> entry and its snapshot remain listed in the API documentation, while <code>gpt-5.2-chat-latest<\/code> has an August 10, 2026 shutdown entry. Before using an integration, verify that your account exposes the exact model identifier you plan to send.<\/p>\n\n\n\n<h3 id=\"is-gpt-5-2-free\" class=\"wp-block-heading\">Is GPT-5.2 free?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">The API page says its Free tier is not supported. If you need ChatGPT access, check the current ChatGPT plan and model picker; if you need developer access, use the documented API model and review the token rates before sending requests. ChatGPT and API billing are separate decisions.<\/p>\n\n\n\n<h3 id=\"what-does-gpt-5-2-cost\" class=\"wp-block-heading\">What does GPT-5.2 cost?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">For the standard API price view checked here: $1.75 per 1,000,000 uncached input tokens, $0.175 per 1,000,000 cached input tokens and $14 per 1,000,000 output tokens. Consumer subscriptions and GlobalGPT use separate billing.<\/p>\n\n\n\n<h3 id=\"was-gpt-5-2-three-times-smarter\" class=\"wp-block-heading\">Was GPT-5.2 three times smarter?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">That broad promise is not supported. The launch report contains strong results on particular evaluations, but the scores vary by task, tool access and model variant. A single percentage cannot establish a general intelligence multiplier.<\/p>\n\n\n\n<h3 id=\"how-should-i-read-the-benchmark-scores\" class=\"wp-block-heading\">How should I read the benchmark scores?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Keep the date, model, metric, tools and evaluation environment attached to every number. Treat the tables as historical evidence about the launch report, not as a current leaderboard or a substitute for testing your own task.<\/p>\n\n\n\n<h2 id=\"next-step\" class=\"wp-block-heading\">Choose the right access path<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Use the <a href=\"https:\/\/developers.openai.com\/api\/docs\/models\/gpt-5.2\">official API documentation<\/a> when you need a model identifier, rate or context limit. Use the <a href=\"https:\/\/help.openai.com\/en\/articles\/6825453-chatgpt-release-notes\">ChatGPT release notes<\/a> for ChatGPT product lifecycle changes. If your project spans research, writing, coding and supporting media, open the <a href=\"https:\/\/www.glbgpt.com\/home\/gpt-6-astra?inviter=hub_content_gpt6astra&amp;login=1\">GPT-6 Astra workspace in GlobalGPT<\/a> and choose the model that fits the next step.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>GPT-5.2 is an OpenAI model for professional work. In ChatGPT, its Instant, Thinking and Pro variants retired on June 12, 2026. The base gpt-5.2 API documentation remains available, and GlobalGPT has a dedicated GPT-5.2 workspace. The right path depends on whether you need ChatGPT, an API identifier or a multi-model workspace. If your work moves [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":6408,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_seopress_robots_primary_cat":"","_seopress_titles_title":"ChatGPT 5.2 Explained: Access, Models, and Benchmarks","_seopress_titles_desc":"GPT-5.2 is now a historical model name. See where ChatGPT access ended, how the API is priced, and what its launch benchmarks actually measured.","_seopress_robots_index":"","footnotes":""},"categories":[7],"tags":[],"class_list":["post-6418","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-chat"],"acf":[],"_links":{"self":[{"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/posts\/6418","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/comments?post=6418"}],"version-history":[{"count":9,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/posts\/6418\/revisions"}],"predecessor-version":[{"id":20379,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/posts\/6418\/revisions\/20379"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/media\/6408"}],"wp:attachment":[{"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/media?parent=6418"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/categories?post=6418"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/wp.glbgpt.com\/wp-json\/wp\/v2\/tags?post=6418"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}