LLM API pricing / OpenAI
What the OpenAI API costs, model by model
Short answer. OpenAI lists 20 models with a per-token price on its API. Input: from $0.10 (GPT-6 Luna) to $21.00 (GPT-5.2 Pro) per 1M tokens. Output: from $0.50 (GPT-6 Luna) to $168.00 (GPT-5.2 Pro). Read on OpenAI's own pages between November 16 and November 17, 2026.
In plain words. You pay for the tokens you send (input) and the tokens the model writes (output). One request of 10,000 tokens in and 1,000 out costs $0.0015 on GPT-6 Luna and $0.378 on GPT-5.2 Pro. Output costs more than input on all 20 models.
Paying for the app? See ChatGPT plans and prices →
Current models
20 models, cheapest input first, verified between November 16, 2026 and November 17, 2026. The last column is the cost of one request of 10,000 input and 1,000 output tokens at these prices, with no cache.
| Model | Input $/1M | Output $/1M | Cached input | Context | Example request | Verified |
|---|---|---|---|---|---|---|
| GPT-6 Luna | $0.10 | $0.50 | $0.01 | 1.1M | $0.0015 | 2026-11-16 |
| GPT-5.6 Luna | $0.20 | $1.20 | $0.02 | 1.1M | $0.0032 | 2026-11-16 |
| GPT-5.4 Mini | $0.75 | $4.50 | $0.075 | 400K | $0.012 | 2026-11-16 |
| GPT-6.1 Terra | $1.00 | $5.00 | $0.05 | 1.1M | $0.015 | 2026-11-16 |
| GPT-5 Search API | $1.25 | $10.00 | $0.125 | — | $0.0225 | 2026-11-16 |
| GPT-5.2 | $1.75 | $14.00 | $0.175 | 400K | $0.0315 | 2026-11-16 |
| GPT-5.6 Terra | $2.00 | $12.00 | $0.20 | 1.1M | $0.032 | 2026-11-16 |
| GPT-6 Sol | $2.00 | $10.00 | $0.20 | 1.1M | $0.03 | 2026-11-16 |
| GPT-6.1 Sol | $2.00 | $10.00 | $0.10 | 1.1M | $0.03 | 2026-11-16 |
| GPT-6.2 Sol | $2.00 | $10.00 | $0.10 | 1.1M | $0.03 | 2026-11-17 |
| GPT-5.6 Sol promo | $4.00 | $20.00 | $0.40 | 1.1M | $0.06 | 2026-11-16 |
| ChatGPT (chat-latest) | $5.00 | $30.00 | $0.50 | 400K | $0.08 | 2026-11-16 |
| GPT-Rosalind | $5.00 | $25.00 | $0.50 | — | $0.075 | 2026-11-16 |
| GPT-Rosalind Discovery | $5.00 | $25.00 | $0.50 | — | $0.075 | 2026-11-16 |
| GPT-6.2 Astra | $8.00 | $40.00 | $0.40 | 1.1M | $0.12 | 2026-11-17 |
| GPT-6 Astra | $10.00 | $50.00 | $1.00 | 1.1M | $0.15 | 2026-11-16 |
| GPT-6 Rosalind promo | $10.00 | $50.00 | $1.00 | — | $0.15 | 2026-11-17 |
| GPT-5.6 Cyber | $12.50 | $75.00 | $1.25 | 400K | $0.20 | 2026-11-16 |
| GPT-6 Cyber | $20.00 | $100.00 | $2.00 | 1.1M | $0.30 | 2026-11-16 |
| GPT-5.2 Pro | $21.00 | $168.00 | — | 400K | $0.378 | 2026-11-16 |
How OpenAI describes its models
- GPT-6 Luna:
GPT-6 Luna is our most efficient model for focused, high-volume tasks.
(developers.openai.com, read September 30, 2026) - GPT-5.6 Luna:
GPT-5.6 Luna is designed for cost-sensitive, high-volume workloads.
(developers.openai.com, read September 30, 2026) - GPT-5.4 Mini:
GPT-5.4 mini brings the strengths of GPT-5.4 to a faster, more efficient model designed for high-volume workloads.
(developers.openai.com, read November 16, 2026) - GPT-6.1 Terra:
A fast, efficient model for visual reasoning, precise computer use, and clear, natural writing.
(developers.openai.com, read October 27, 2026) - GPT-5.2:
GPT-5.2 is our previous flagship model for complex professional work.
(developers.openai.com, read October 26, 2026) - GPT-5.6 Terra:
GPT-5.6 Terra is designed for workloads that balance intelligence and cost.
(developers.openai.com, read September 30, 2026) - GPT-6 Sol:
GPT-6 Sol is built for complex coding and agentic workflows.
(developers.openai.com, read September 30, 2026) - GPT-6.1 Sol:
Near-Astra performance for complex work at a lower cost.
(developers.openai.com, read November 3, 2026) - GPT-6.2 Sol:
GPT-6.2 Sol for everyday coding and agent workflows.
(developers.openai.com, read November 17, 2026) - ChatGPT (chat-latest):
chat-latest points to the latest Instant model currently used in ChatGPT. We recommend leveraging GPT-6 Astra for production API usage.
(developers.openai.com, read September 30, 2026) - GPT-6.2 Astra:
Use GPT-6.2 Astra for demanding reasoning and coding tasks
(developers.openai.com, read November 17, 2026) - GPT-6 Astra:
GPT-6 Astra is our most capable model for the most demanding work.
(developers.openai.com, read September 30, 2026) - GPT-5.6 Cyber:
Cybersecurity model for authorized vulnerability research and security testing.
(developers.openai.com, read November 6, 2026) - GPT-6 Cyber:
Investigate a vulnerability, reproduce it in an authorized test environment, or verify a fix.
(developers.openai.com, read November 6, 2026) - GPT-5.2 Pro:
GPT-5.2 Pro is our previous pro model for complex professional work.
(developers.openai.com, read October 26, 2026)
How OpenAI prices its models
- 19 of the 20 models publish a cached-input price, at 5% or 10% of the input price.
- 11 models charge more for long prompts (prompts over 272,000 tokens): GPT-6 Luna $0.20 / $0.75, GPT-5.6 Luna $0.40 / $1.80, GPT-6.1 Terra $2.00 / $7.50, GPT-5.6 Terra $4.00 / $18.00, GPT-6 Sol $4.00 / $15.00, GPT-6.1 Sol $4.00 / $15.00, GPT-6.2 Sol $4.00 / $15.00, GPT-5.6 Sol $8.00 / $30.00, GPT-6.2 Astra $16.00 / $60.00, GPT-6 Astra $20.00 / $75.00, GPT-6 Rosalind $20.00 / $75.00.
- 2 prices are promotional: GPT-5.6 Sol until November 30, 2026, no list price published for after; GPT-6 Rosalind at least through December 31, 2026, no list price published for after.
Second price lists
Prices OpenAI prints for other ways of buying the same tokens, per 1M tokens, only where its page prints a figure for the model.
Batch · half of Standard on every column
- GPT-6 Luna: $0.05 input, $0.005 cached input, $0.0625 cache write, $0.25 output (read September 24, 2026)
- GPT-5.6 Luna: $0.10 input, $0.01 cached input, $0.125 cache write, $0.60 output (read October 26, 2026)
- GPT-5.4 Mini: $0.375 input, $0.0375 cached input, $2.25 output (read October 26, 2026)
- GPT-6.1 Terra: $0.50 input, $0.025 cached input, $0.625 cache write, $2.50 output (read October 27, 2026)
- GPT-5.2: $0.875 input, $0.0875 cached input, $7.00 output (read October 26, 2026)
- GPT-5.6 Terra: $1.00 input, $0.10 cached input, $1.25 cache write, $6.00 output (read October 26, 2026)
- GPT-6 Sol: $1.00 input, $0.10 cached input, $1.25 cache write, $5.00 output (read October 26, 2026)
- GPT-6.1 Sol: $1.00 input, $0.05 cached input, $1.25 cache write, $5.00 output (read September 30, 2026)
- GPT-6.2 Sol: $1.00 input, $0.05 cached input, $1.25 cache write, $5.00 output (read November 17, 2026)
- GPT-5.6 Sol: $2.00 input, $0.20 cached input, $2.50 cache write, $10.00 output (read September 16, 2026)
- GPT-6.2 Astra: $4.00 input, $0.20 cached input, $5.00 cache write, $20.00 output (read November 17, 2026)
- GPT-6 Astra: $5.00 input, $0.50 cached input, $6.25 cache write, $25.00 output (read September 16, 2026)
- GPT-5.2 Pro: $10.50 input, $84.00 output (read October 26, 2026)
OpenAI: "Learn how to use OpenAI's Batch API to send asynchronous groups of requests with 50% lower costs, a separate pool of significantly higher rate limits, and a clear 24-hour turnaround time." source
Flex · half of Standard on every column, the same figures as Batch
- GPT-6 Luna: $0.05 input, $0.005 cached input, $0.0625 cache write, $0.25 output (read September 24, 2026)
- GPT-5.6 Luna: $0.10 input, $0.01 cached input, $0.125 cache write, $0.60 output (read October 26, 2026)
- GPT-5.4 Mini: $0.375 input, $0.0375 cached input, $2.25 output (read October 26, 2026)
- GPT-6.1 Terra: $0.50 input, $0.025 cached input, $0.625 cache write, $2.50 output (read October 27, 2026)
- GPT-5.2: $0.875 input, $0.0875 cached input, $7.00 output (read October 26, 2026)
- GPT-5.6 Terra: $1.00 input, $0.10 cached input, $1.25 cache write, $6.00 output (read October 26, 2026)
- GPT-6 Sol: $1.00 input, $0.10 cached input, $1.25 cache write, $5.00 output (read October 26, 2026)
- GPT-6.1 Sol: $1.00 input, $0.05 cached input, $1.25 cache write, $5.00 output (read September 30, 2026)
- GPT-5.6 Sol: $2.00 input, $0.20 cached input, $2.50 cache write, $10.00 output (read September 16, 2026)
- GPT-6 Astra: $5.00 input, $0.50 cached input, $6.25 cache write, $25.00 output (read September 16, 2026)
OpenAI: "Flex processing provides lower costs for Responses or Chat Completions requests in exchange for slower response times and occasional resource unavailability." source
Fast mode · double Standard on every column, except the GPT-4.1 and GPT-4o models, which the page prices at less than double
- GPT-6 Luna: $0.20 input, $0.02 cached input, $0.25 cache write, $1.00 output (read September 24, 2026)
- GPT-5.6 Luna: $0.40 input, $0.04 cached input, $0.50 cache write, $2.40 output (read October 26, 2026)
- GPT-5.4 Mini: $1.50 input, $0.15 cached input, $9.00 output (read October 26, 2026)
- GPT-6.1 Terra: $2.00 input, $0.10 cached input, $2.50 cache write, $10.00 output (read October 27, 2026)
- GPT-5.2: $3.50 input, $0.35 cached input, $28.00 output (read October 26, 2026)
- GPT-5.6 Terra: $4.00 input, $0.40 cached input, $5.00 cache write, $24.00 output (read October 26, 2026)
- GPT-6 Sol: $4.00 input, $0.40 cached input, $5.00 cache write, $20.00 output (read October 26, 2026)
- GPT-6.1 Sol: $4.00 input, $0.20 cached input, $5.00 cache write, $20.00 output (read September 30, 2026)
- GPT-6.2 Sol: $4.00 input, $0.20 cached input, $5.00 cache write, $20.00 output (read November 17, 2026)
- GPT-5.6 Sol: $8.00 input, $0.80 cached input, $10.00 cache write, $40.00 output (read September 16, 2026)
- GPT-6.2 Astra: $16.00 input, $0.80 cached input, $20.00 cache write, $80.00 output (read November 17, 2026)
- GPT-6 Astra: $20.00 input, $2.00 cached input, $25.00 cache write, $100.00 output (read September 16, 2026)
OpenAI: "Priority processing was renamed Fast mode on July 30, 2026." source
Ultrafast · the figures printed on the pricing page's Ultrafast tab
- GPT-6.1 Sol: $12.00 input, $0.60 cached input, $15.00 cache write, $60.00 output (read November 17, 2026)
- GPT-6.2 Sol: $6.00 input, $0.30 cached input, $7.50 cache write, $30.00 output (read November 17, 2026)
- GPT-6.2 Astra: $24.00 input, $1.20 cached input, $30.00 cache write, $120.00 output (read November 17, 2026)
- GPT-6 Astra: $60.00 input, $6.00 cached input, $75.00 cache write, $300.00 output (read November 17, 2026)
OpenAI: "Ultrafast mode is the fastest service tier in the OpenAI API. It supports GPT-6.2 Astra, GPT-6.2 Sol, GPT-6 Astra, and GPT-6.1 Sol. Use it when speed justifies the higher cost." source
Retirements
Deprecated, still callable: GPT-5 (shutdown December 11, 2026); GPT-5 mini (shutdown December 11, 2026); GPT-5 nano (shutdown December 11, 2026); GPT-5 Pro (shutdown December 11, 2026); o3 (shutdown December 11, 2026); o3-pro (shutdown December 11, 2026); GPT-4.1 (shutdown January 14, 2027); GPT-4.1 Mini (shutdown January 14, 2027); GPT-4o (shutdown January 14, 2027); GPT-4o Mini (shutdown January 14, 2027); GPT-5.5 Cyber (shutdown February 1, 2027); GPT-5.1 (shutdown April 1, 2027); GPT-5.3 Codex (shutdown April 1, 2027); GPT-5.4 nano (shutdown April 1, 2027).
Retired: GPT-3.5 Turbo (October 23, 2026), GPT-4 (October 23, 2026), GPT-4 Turbo (October 23, 2026), o1 (October 23, 2026), o1-pro (October 23, 2026), o3-mini (October 23, 2026).
- October 23, 2026: Shutdown of o1, o1-pro, o3-mini, GPT-4, GPT-4 Turbo and GPT-3.5 Turbo source
- December 11, 2026 (announced): Shutdown of GPT-5, GPT-5 mini, GPT-5 nano, GPT-5 Pro, o3 and o3-pro source
- January 14, 2027 (announced): Shutdown of GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini (notice date October 14, 2026; first recorded by this site on November 16, 2026) source
- February 1, 2027 (announced): Shutdown of GPT-5.5 Cyber, listed as gpt-5.5-cyber-preview (notice date November 5, 2026; OpenAI names no replacement model) source
- April 1, 2027 (announced): Shutdown of GPT-5.3 Codex, GPT-5.1 and GPT-5.4 nano source
Recorded price changes
- November 17, 2026 · price down 1.5×. OpenAI cut the price of GPT-6 Rosalind, a restricted life-sciences model, from $15.00 to $10.00 per million input tokens (15.00 ÷ 10.00 = 1.5×) and from $75.00 to $50.00 for output (1.5×), and made it promotional. The last reading at the old price was the weekly check of November 16, 2026, 'gpt-6-rosalind | $15.00 | $75.00 | $1.50 | $18.75 | $37.50 |'; on November 17, 2026 the Markdown version of the pricing page printed 'gpt-6-rosalind (short context) | $10.00 | $50.00 | $1.00 | $12.50 | $25.00 |' and 'gpt-6-rosalind (long context) | $20.00 | $75.00 | $2.00 | $25.00 | $50.00 |', with 'GPT-6 Astra for Law, GPT-6.2 Astra for Law, and GPT-6 Rosalind promotional pricing is available at least through December 31, 2026.' 'At least through' is not a firm end date, and OpenAI publishes no list price for after the promotion. The same paragraph adds: 'For standalone FedRAMP endpoints, GPT-6 Rosalind rates are 10% above the corresponding global rates.' OpenAI's changelog does not date the change, so the date here is the first day it was seen. source
- November 16, 2026 · correction of our own reading. Correction of AI Signal's record, not a price change by OpenAI: no price moved. Until November 16, 2026 AI Signal listed GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini as deprecated with no shutdown date. OpenAI's deprecations page had listed 'Jan 14, 2027' for the four since its notice of October 14, 2026; AI Signal's reader missed the page's new table format. Read on November 16, 2026, the page prints 'Jan 14, 2027 | GPT-4o ( gpt-4o-2024-08-06 , gpt-4o-2024-11-20 ) | Deprecated | gpt-6-luna | Oct 14, 2026 |', and the same date, status, replacement and notice date for 'GPT-4o mini ( gpt-4o-mini-2024-07-18 )', 'GPT-4.1 ( gpt-4.1-2025-04-14 )' and 'GPT-4.1 mini ( gpt-4.1-mini-2025-04-14 )'. The four rows now carry that shutdown date and GPT-6 Luna as the recommended replacement. source
- November 9, 2026 · price terms changed, figure unchanged, see below. Change of the promotion's end date, not of the price. OpenAI's pricing page and GPT-5.6 Sol's model page, read on November 9, 2026, say 'GPT-5.6 Sol’s promotional pricing is available until November 30, 2026.' In every weekly reading from October 5 to November 2, 2026 the pricing page said 'GPT-5.6 Sol’s promotional pricing is available at least through November 21, 2026.' The promotional price, $4.00 input and $20.00 output per million tokens, is unchanged, and OpenAI still publishes no list price for after it. First seen by AI Signal on November 9, 2026; OpenAI does not say when it changed the wording. source
- October 29, 2026 · correction of our own reading. Correction of our own reading, not a price change by OpenAI. The Markdown version of OpenAI's pricing page (developers.openai.com/api/docs/pricing.md) prints, in the Standard table of GPT-5.6 Sol and the Cyber models, the row '| gpt-5.6-cyber | $12.50 | $75.00 | $1.25 | $15.625 | $31.25 | - | - | - | - | - |': a 30-minute cache write of $15.625 and a 12-hour cache write of $31.25 per million tokens, read on October 29, 2026. Before that this dataset showed GPT-5.6 Cyber with no cache write price; input ($12.50), cached input ($1.25) and output ($75.00) are unchanged. When OpenAI began printing these figures is not known, so the date here is the day they were read. source
- October 26, 2026 · correction of our own reading. Correction of our own reading, not a price change by OpenAI. The entry of October 5, 2026 (openai-deprecated-model-pages-price-found) said that OpenAI's pricing page no longer listed GPT-4, GPT-4 Turbo, GPT-3.5 Turbo, o1, o1-pro and o3-mini. That was AI Signal's reading of the page, which did not see its folded 'All models' table: the page draws it in the browser, and its Markdown version (developers.openai.com/api/docs/pricing.md) prints it. Read there on October 26, 2026, three days after the six were shut down, the Standard table still lists them, at the same prices AI Signal carries: 'gpt-4-0613 | $30.00 | $60.00 | -', 'gpt-4-turbo-2024-04-09 | $10.00 | $30.00 | -', 'gpt-3.5-turbo-0125 | $0.50 | $1.50 | -', 'o1 | $15.00 | $60.00 | $7.50', 'o1-pro | $150.00 | $600.00 | -' and 'o3-mini | $1.10 | $4.40 | $0.55' (Input, Output, Cached input, per 1M tokens). No figure shown here changes. Whether the table listed them on October 5, 2026 is not known. source
- October 26, 2026 · correction of our own reading. Correction of our own reading, not a price change by OpenAI. Since September 3, 2026 AI Signal showed GPT-5, GPT-5 mini and o3 with no price, and said GPT-5's price was no longer published. OpenAI's pricing page does print them, in its folded 'All models' table, which the page draws in the browser and its Markdown version (developers.openai.com/api/docs/pricing.md) prints. Read there on October 26, 2026, the Standard table prints 'gpt-5 | $1.25 | $10.00 | $0.125', 'gpt-5-mini | $0.25 | $2.00 | $0.025' and 'o3 | $2.00 | $8.00 | $0.50' (Input, Output, Cached input, per 1M tokens); each model page, read the same day, prints the same figures. All three are deprecated and OpenAI shuts them down on December 11, 2026. source
- October 5, 2026 · correction of our own reading. Correction of our own reading, not a price change by OpenAI. Until October 5, 2026 AI Signal showed GPT-4, GPT-4 Turbo, GPT-3.5 Turbo, o1, o1-pro and o3-mini with no price, because it read OpenAI's pricing page, which no longer lists them, and not their model pages, which still do. Read by hand on October 5, 2026, with the page's 'Batch API price' switch off (Standard prices), the model pages print, per 1M tokens: GPT-4 'Input $30.00 Output $60.00'; GPT-4 Turbo 'Input $10.00 Output $30.00'; GPT-3.5 Turbo 'Input $0.50 Output $1.50'; o1 'Input $15.00 Cached input $7.50 Output $60.00'; o1-pro 'Input $150.00 Output $600.00'; o3-mini 'Input $1.10 Cached input $0.55 Output $4.40'. All six are deprecated and OpenAI shuts them down on October 23, 2026; each row now cites its model page. source
- September 16, 2026 · added to this dataset, our omission: GPT-6 Astra, GPT-Rosalind. Full explanation
- September 10, 2026 · correction of our own reading · figure shown here up 2.0×. Correction of our own reading, not a price change by OpenAI. From September 3 to September 10, 2026 AI Signal published GPT-5.6 Sol at $2 / $10, Terra at $1 / $6 and Luna at $0.10 / $0.60 per million tokens, read from the Batch tab of OpenAI's pricing page, which is always 50% of Standard (the page's four tabs, Standard, Batch, Flex and Fast mode, look identical in plain text), and labelled as a promotion ending November 21 with double the price after it. The Standard short-context prices, verified by hand on September 10, are Sol $4 / $20, Terra $2 / $12 and Luna $0.20 / $1.20, and the published figures doubled to them. The November 21 date was OpenAI's, but it belongs to Sol alone and is not an end date: OpenAI states that Sol's $4 / $20 is promotional pricing, a 20% reduction on input and 33% on output, available at least through November 21, 2026, and publishes no list price for after it; Terra and Luna carry no promotion. On September 10 this entry wrongly said there was no promotion to end; corrected on September 15, when the Sol row got its promotion back, marked as at least through that date, without an invented list price. source
Analysis that cites OpenAI models
- Prompt Caching Is Not 10% Everywhere: What a Cache Hit Costs Across Eight Providers · names 8 OpenAI models
- OpenAI Shuts Down GPT-4, GPT-4 Turbo, GPT-3.5 Turbo, o1, o1-pro and o3-mini on October 23, 2026: the Replacements and What They Cost · names 7 OpenAI models
- GPT-6.2 Astra at $8 / $40, GPT-6.2 Sol at $2 / $10 · names 5 OpenAI models
- AI Pricing This Week: Large 4 Cut, New Hosts, Oct 26-Nov 1 · names 5 OpenAI models
- AI Pricing This Week: Claude Ultra, Gemini 2.5, Nov 2-8 · names 3 OpenAI models
- GPT-6.1 Terra API Pricing: $1 In, $5 Out, Half of GPT-6.1 Sol · names 3 OpenAI models
- AI Pricing This Week: OpenAI Shutdown, Mistral, Oct 19-25 · names 3 OpenAI models
- Same List Price, Different Bill: Five Pairs of LLM APIs Where the Workload Decides What You Pay · names 3 OpenAI models
- OpenAI's Pricing Page Has Four Tabs That Look the Same in Plain Text. We Published the Wrong One for Seven Days. · names 3 OpenAI models
- AI Pricing This Week: Fable 5.2, Gemini 4 Carbon, Nov 9-15 · names 2 OpenAI models
- Claude Sonnet 4.5 Retires on November 30, 2026: What Moving to Claude Sonnet 5.5 Does to the Bill · names 2 OpenAI models
- The Second Price List: What Batch, Flex, Priority and Data-Sharing Discounts Cost Across Seven Providers · names 2 OpenAI models
- OpenAI Shuts Down GPT-4o and GPT-4.1 on January 14, 2027 · names 1 OpenAI model
- Qwen4-Max: $1.50 / $4.50 per Million Tokens, Under Qwen3.8-Max, With Cached Input 3.3× Cheaper · names 1 OpenAI model
- What Changes on December 1, 2026: Gemini 4 Argon and Carbon Double, GPT-5.6 Sol's Promotion Ends, Gemini 2.5 Shuts Down · names 1 OpenAI model
- AI Pricing This Week: Gemini 4 Argon, Muse Wonder, Oct 12-18 · names 1 OpenAI model
- Five Things Decide What You Pay for a Model, and Only One of Them Is the Model · names 1 OpenAI model
- Claude Code vs OpenAI Codex: Which Plan Includes What · October 20, 2026
FAQ
What is the cheapest OpenAI model on the API?
GPT-6 Luna, at $0.10 per 1M input tokens and $0.50 per 1M output tokens, as OpenAI's pricing page printed it on November 16, 2026.
How we checked
- Each price was read on OpenAI's pricing page (developers.openai.com) or on a model page it links. The Verified column links to that page and gives the day.
- Every Monday a script reads OpenAI's pricing page again and compares each row with the figures here. A difference is read by hand before a price changes.
- Quotes are copied as OpenAI's pages print them, with the day they were read. A script checks every week that the model descriptions are still on their pages.
- The example request is worked out from the per-1M prices: 10,000 input tokens and 1,000 output tokens, no cache, no batch discount, no long-prompt tier.
Full method: methodology. All providers: the full pricing table.