{"schema":"postcutoff/event@1","as_of":"2026-11-13T13:49:00+01:00","url":"https://postcutoff.com/e/2026-09-29-chatgpt-pro-500-ultrafast/","md":"https://postcutoff.com/e/2026-09-29-chatgpt-pro-500-ultrafast/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-09-29-chatgpt-pro-500-ultrafast","date":"2026-09-29","date_precision":"day","short_title":"OpenAI adds a $500/month Pro 500 plan and the Ultrafast speed tier, and halves the $200 Pro allowance","deck":null,"takeaway":"At DevDay on Sept 29, 2026, OpenAI launched \"Ultrafast\", a premium speed tier (up to 8x faster token generation, about 300 tokens/s, in Codex and up to 6x in the API, at 6x the API price), and \"Pro 500\", a $500/month ChatGPT plan with 25x the Plus allowance and Ultrafast.","category":"business","category_label":"Business","importance":3,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":[{"n":1,"title":"OpenAI: DevDay 2026 Recap (Ultrafast, Pro 500)","url":"https://openai.com/index/devday-2026-recap/","type":"official","group":"primary","domain":"openai.com"},{"n":2,"title":"OpenAI Help: About ChatGPT Pro tiers","url":"https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers","type":"docs","group":"primary","domain":"help.openai.com"},{"n":3,"title":"OpenAI API docs: Ultrafast mode","url":"https://developers.openai.com/api/docs/guides/ultrafast-mode","type":"docs","group":"primary","domain":"developers.openai.com"},{"n":4,"title":"OpenAI API pricing: Ultrafast table","url":"https://developers.openai.com/api/docs/pricing?latest-pricing=ultrafast","type":"docs","group":"primary","domain":"developers.openai.com"},{"n":5,"title":"OpenAI: Previewing Ultrafast mode: GPT-5.6 Sol at up to 14X the speed (Aug 13, 2026)","url":"https://openai.com/index/previewing-ultrafast/","type":"official","group":"primary","domain":"openai.com"},{"n":6,"title":"ChatGPT docs: agent speed configuration","url":"https://learn.chatgpt.com/docs/agent-configuration/speed","type":"docs","group":"primary","domain":"learn.chatgpt.com"},{"n":7,"title":"Thibault Sottiaux on X: Pro $200 usage change","url":"https://x.com/thsottiaux/status/2104823812042940713","type":"official","group":"primary","domain":"x.com"},{"n":8,"title":"Thibault Sottiaux on X: new plan multipliers","url":"https://x.com/thsottiaux/status/2104951965184925941","type":"official","group":"primary","domain":"x.com"},{"n":9,"title":"NVIDIA: How NVIDIA GPUs help accelerate OpenAI's GPT-6 Astra Ultrafast (Oct 1, 2026)","url":"https://blogs.nvidia.com/blog/gpus-openai-gpt-6-astra-ultrafast/","type":"official","group":"primary","domain":"blogs.nvidia.com"},{"n":10,"title":"OpenAI: A model guide for the GPT-6 family (Oct 2, 2026)","url":"https://openai.com/index/practical-guide-building-gpt-6","type":"official","group":"primary","domain":"openai.com"},{"n":11,"title":"CNBC: Cerebras stock pops 9% after Sam Altman calls the chipmaker a 'close partner' (Oct 5)","url":"https://www.cnbc.com/2026/10/05/cerebras-cbrs-sam-altman-close-partner.html","type":"press","group":"press","domain":"cnbc.com"},{"n":12,"title":"Engadget: OpenAI adds $500(!) Pro subscription, nerfs its existing $200 tier","url":"https://www.engadget.com/2272106/openai-adds-dollar500-pro-subscription-nerfs-its-existing-dollar200-tier/","type":"press","group":"press","domain":"engadget.com"},{"n":13,"title":"The New Stack: OpenAI halves $200 plan allowance, launches $500 plan","url":"https://thenewstack.io/openai-halves-200-plan/","type":"press","group":"press","domain":"thenewstack.io"},{"n":14,"title":"The Decoder: OpenAI reopens its $200 Pro plan but cuts API credits in half","url":"https://the-decoder.com/openai-reopens-its-200-pro-plan-but-cuts-api-credits-in-half-as-it-nudges-users-toward-pay-per-use/","type":"press","group":"press","domain":"the-decoder.com"},{"n":15,"title":"The Decoder: Ultrafast pricing ($60/$300 for Astra)","url":"https://the-decoder.com/openai-expands-codex-and-its-api-at-devday-with-security-scans-a-decisions-api-and-ultrafast/","type":"press","group":"press","domain":"the-decoder.com"},{"n":16,"title":"CNBC live blog: OpenAI launching new Pro tier","url":"https://www.cnbc.com/2026/09/29/openai-devday-2026-live-updates.html","type":"press","group":"press","domain":"cnbc.com"},{"n":17,"title":"CNET: Everything announced at OpenAI DevDay","url":"https://www.cnet.com/tech/services-and-software/everything-announced-at-openai-devday-subscription-changes-new-models-and-dots/","type":"press","group":"press","domain":"cnet.com"},{"n":18,"title":"mixed-news: OpenAI says its own models optimised Astra Ultrafast's inference on Nvidia GPUs","url":"https://mixed-news.com/en/openai-internal-models-optimised-astra-ultrafast-nvidia-blackwell/","type":"press","group":"press","domain":"mixed-news.com"}],"official":10,"filed":"2026-09-29","updated":"2026-10-05","orgs":["OpenAI"],"title":"OpenAI adds a $500/month Pro 500 plan and the Ultrafast speed tier, and halves the $200 Pro allowance","summary":"At DevDay on Sept 29, 2026, OpenAI launched \"Ultrafast\", a premium speed tier (up to 8x faster token generation, about 300 tokens/s, in Codex and up to 6x in the API, at 6x the API price), and \"Pro 500\", a $500/month ChatGPT plan with 25x the Plus allowance and Ultrafast. It also reopened the $200 Pro plan to new subscribers but cut its ChatGPT Work/Codex allowance from 20x to 10x Plus and its GPT-6 Pro messages from 200 to 100 a week. OpenAI's Thibault Sottiaux framed the change as moving subscriptions toward API-equivalent value.","key_facts":["Pro 500: $500/month, 'our highest usage allowance at 25 times the ChatGPT Plus allowance', includes Ultrafast; available now","Ultrafast: up to 8x faster (300 tokens/s) in Codex and up to 6x in the API; GPT-6 Astra Ultrafast available in the API (all users, low rate limits) and in ChatGPT Work/Codex on Pro 500 and Enterprise; GPT-6.1 Sol Ultrafast 'coming soon'","Ultrafast API price for gpt-6-astra: $60 input / $6 cached input / $75 cache writes / $300 output per 1M tokens (≤272K context); $120/$12/$150/$450 above 272K; i.e. 6x standard","API usage: set service_tier: \"ultrafast\" with model gpt-6-astra; WebSockets recommended; default Ultrafast limits 500K TPM (tiers 1–3), 1M (tier 4), 5M (tier 5); preview access for GPT-5.6 Sol via account teams","New plan multipliers (Sottiaux): Plus = 1x, Pro 100 = 5x, Pro 200 = 10x (was 20x); Pro 500 = 25x","Pro 200 changes: GPT-6 Pro messages in Chat cut from 200 to 100/week; existing subscribers keep previous allowance through Oct 29, 2026, then get a one-time grant of 62,500 credits (worth $2,500 per The New Stack) expiring Dec 31, 2026; the 5-hour limit will not return","Sottiaux (Sept 29, 12.8M views): 'if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan'","In ChatGPT, Ultrafast draws from included usage first, then from purchased credits","Ultrafast first appeared on Aug 13, 2026 as an API preview for GPT-5.6 Sol ('up to 14X the speed', up to 750 output tokens/s, powered by Cerebras); DevDay made it broadly available for GPT-6 Astra","Hardware (NVIDIA, Oct 1 2026): GPT-6 Astra Ultrafast runs on NVIDIA Blackwell GPUs; OpenAI says it used its own models to optimize inference on them ('Astra can turn that knowledge into high-performance kernels', Philippe Tillet). The Aug 2026 GPT-5.6 Sol preview had run on Cerebras","As of Oct 2, 2026, OpenAI's GPT-6 model guide still lists Ultrafast as 'available for GPT-6 Astra' only; GPT-6.1 Sol Ultrafast not yet shipped","Pre-event reports had described a $500/month 'ChatGPT Pro Max' plan; it shipped as 'Pro 500'","NVIDIA blog (Oct 1): GPT-6 Astra Ultrafast runs on Blackwell GPUs. OpenAI compute CTO Uday Ruddarraju: 'We used our internal models to optimize inference on NVIDIA GPUs, and NVIDIA's programmability helped us deliver the acceleration behind Astra Ultrafast'; inference lead Philippe Tillet: NVIDIA's tooling 'enabled us to make our models exceptionally good at programming Blackwell and Rubin GPUs'","Cerebras fallout (CNBC, Oct 5): Cerebras stock fell 20% to its lowest level the week of Sept 28 after reports that GPT-6.1 Sol Ultrafast would run on NVIDIA GPUs rather than Cerebras chips. On Friday Oct 2 Altman posted on X: 'There is some speculation about our partnership with Cerebras. Cerebras is a close partner, and we have a deep engagement pushing on the frontiers of speed.' The stock rose ~3% after hours and ~9% on Monday Oct 5; market cap ~$43B vs $95B after its May IPO. Citi: labs' newest models 'would initially roll out on internal chips before running on third-party or Cerebras cloud'"],"key_numbers":[],"tags":["pricing","subscriptions","chatgpt-pro","ultrafast","inference-speed","codex","api","devday","openai"],"science":null,"body_md":"## What happened\n\nOpenAI changed its ChatGPT subscription tiers on DevDay:\n- **Pro 500 ($500/month)** is the new top consumer plan. It gives 25x the Plus allowance and access to **Ultrafast**. A $500 plan had been\n  reported before the event under the name \"Pro Max\".\n- **Pro 200** reopened to new subscribers after a pause, but at half its former relative allowance (10x Plus instead of 20x). GPT-6 Pro chat\n  messages fall from 200 to 100 a week. Subscribers who were active at the cutoff keep their old allowance through Oct 29, 2026. After that,\n  The New Stack reports they receive a one-time grant of 62,500 credits that expires Dec 31, 2026.\n- **Pro 100** stays at 5x. All Pro tiers include one dot, and dot conversations don't draw on usage.\n\nSottiaux (head of product and platform, per The New Stack) announced the Pro 200 change hours before the keynote. That post reached about 12.8M views. He said the\nnew usage \"will net out at half the dollar in API spend compared to the old Pro $200 plan\". He argued that cheaper models (GPT-6 Sol and\nLuna at half price) mean subscribers still get more work done than a month earlier. He said OpenAI will not reintroduce the 5-hour limit,\nand that over time API prices should fall so far that \"it makes sense for most to buy usage as needed\". Press and developer reaction to\nthe cut was largely negative (The New Stack, Engadget, CNET).\n\n**Ultrafast** is a new API service tier (`service_tier: \"ultrafast\"`). It is broadly available for GPT-6 Astra at 6x standard prices\n($60/$300 per 1M input/output tokens). GPT-5.6 Sol has preview access, and GPT-6.1 Sol is \"coming soon\". OpenAI's existing \"Fast mode\"\ncosts 2x standard.\n\n## Why it matters\n\nFrontier labs have used a $200/month top tier since late 2024, and the 20x-Plus allowance was the norm among them. OpenAI broke that\npattern in both directions: it cut the $200 plan's allowance and added a $500 tier. It also sells speed as its own product. The stated\naim is to bring flat-rate subscriptions in line with pay-per-use API pricing.","disputed":[],"related":[{"id":"2026-09-29-openai-devday-2026","url":"https://postcutoff.com/e/2026-09-29-openai-devday-2026/","date":"2026-09-29","date_precision":"day","short_title":"OpenAI DevDay 2026 brings dots agents, GPT-6.1 Sol, Ultrafast and a $500 Pro plan","deck":null,"takeaway":"DevDay marks OpenAI's shift from chat and coding tools toward persistent, proactive agents (dots), and toward ChatGPT as a workspace platform (Space, Pages, plugin extensions, Marketplace).","category":"product","category_label":"Products","importance":4,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":41,"official":23,"filed":"2026-09-29","updated":"2026-10-18","orgs":["OpenAI"]},{"id":"2026-09-29-gpt-6-1-sol","url":"https://postcutoff.com/e/2026-09-29-gpt-6-1-sol/","date":"2026-09-29","date_precision":"day","short_title":"OpenAI releases GPT-6.1 Sol","deck":"Near-Astra performance at one-fifth of Astra's price","takeaway":"Near-flagship capability is now available at mid-tier prices, a week after the previous mid-tier model.","category":"model-release","category_label":"Model releases","importance":4,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":15,"official":11,"filed":"2026-09-29","updated":"2026-10-10","orgs":["OpenAI"]},{"id":"2026-09-29-openai-dots","url":"https://postcutoff.com/e/2026-09-29-openai-dots/","date":"2026-09-29","date_precision":"day","short_title":"OpenAI launches dots, always-on personal agents powered by GPT-6 Astra","deck":null,"takeaway":"Dots are OpenAI's first mass-market product built around a persistent agent that acts without being prompted, rather than a chat or coding session.","category":"agents","category_label":"Agents","importance":4,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":26,"official":13,"filed":"2026-09-29","updated":"2026-10-09","orgs":["OpenAI"]},{"id":"2026-09-03-gpt-6-astra","url":"https://postcutoff.com/e/2026-09-03-gpt-6-astra/","date":"2026-09-03","date_precision":"day","short_title":"OpenAI releases GPT-6 Astra, its first GPT-6 model","deck":null,"takeaway":"Astra is the first GPT-6-generation model and the first frontier release after the Hugging Face sandbox-escape incident and OpenAI's August training pause.","category":"model-release","category_label":"Model releases","importance":5,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":20,"official":11,"filed":"2026-09-29","updated":"2026-10-19","orgs":["OpenAI"]},{"id":"2026-07-30-gpt-5-6-price-cut","url":"https://postcutoff.com/e/2026-07-30-gpt-5-6-price-cut/","date":"2026-07-30","date_precision":"day","short_title":"OpenAI cuts GPT-5.6 Luna price 80% and Terra 20%","deck":null,"takeaway":"Evidence of rapid commoditization of the \"utility\" tier of frontier-lab models in 2026; it foreshadowed the further 50% cut with GPT-6 Sol/Luna in September.","category":"business","category_label":"Business","importance":2,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":6,"official":4,"filed":"2026-09-29","updated":"2026-09-29","orgs":["OpenAI"]}],"people":[{"id":"sam-altman","name":"Sam Altman","url":"https://postcutoff.com/person/sam-altman/"},{"id":"thibault-sottiaux","name":"Thibault Sottiaux","url":"https://postcutoff.com/person/thibault-sottiaux/"},{"id":"uday-ruddarraju","name":"Uday Ruddarraju","url":"https://postcutoff.com/person/uday-ruddarraju/"}],"posts":[{"id":"2026-09-29-thsottiaux-pro-200-usage-change","title":"Sottiaux: Pro $200 reopens with usage that nets out at half the API-dollar value","url":"https://postcutoff.com/p/2026-09-29-thsottiaux-pro-200-usage-change/"},{"id":"2026-09-29-thsottiaux-plan-multipliers","title":"Sottiaux: new plan multipliers Plus 1x, Pro 100 5x, Pro 200 10x","url":"https://postcutoff.com/p/2026-09-29-thsottiaux-plan-multipliers/"}],"videos":[{"id":"openai-devday-2026-keynote","title":"Live from OpenAI DevDay 2026: Keynote","url":"https://postcutoff.com/v/openai-devday-2026-keynote/"},{"id":"openai-build-faster-with-ultrafast","title":"Build faster with Ultrafast","url":"https://postcutoff.com/v/openai-build-faster-with-ultrafast/"}],"models":[],"changes":[{"date":"2026-10-05","type":"edit","text":"Added the Cerebras share-price fallout and Altman's 'close partner' post (CNBC)"},{"date":"2026-10-04","type":"edit","text":"Added OpenAI executives' quotes from NVIDIA's Oct 1 blog (internal models optimised Ultrafast inference)"},{"date":"2026-10-02","type":"edit","text":"Added NVIDIA Blackwell hardware detail (NVIDIA blog, Oct 1) and the Oct 2 GPT-6 model guide status of 6.1 Sol Ultrafast."},{"date":"2026-09-29","type":"filed","text":"Created"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-09-29","run":null,"sources_read":null,"updated":"2026-10-05","human_review":null,"version":{"date":"2026-10-06"}},"gaps":[{"model_id":"gpt-6-astra","name":"GPT-6 Astra","cutoff":"2026-04","days_after":152,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":91,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":182,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":121,"in_training_data":false}],"short_url":null}