GPT 6 pricing: API costs, Pro plans and Codex
GPT 6 pricing separates API token costs from ChatGPT plans and Codex credits. Compare models, caching and speed modes before estimating monthly usage.
GPT-6 Astra, Sol and Luna API rates
| Model | Input | Cached read | Cache write | Output |
|---|---|---|---|---|
| GPT-6 Astra | $10.00 | $1.00 | $12.50 | $50.00 |
| GPT-6 Sol | $2.00 | $0.20 | $2.50 | $10.00 |
| GPT-6 Luna | $0.10 | $0.01 | $0.125 | $0.50 |
All three share a 1,050,000-token context window and a 128,000-token maximum output. Output uses part of that context window.
Astra service modes
USD per million text tokens. Last verified: 2026-09-24. All tables and examples use the calculator’s model configuration.
| Mode | Input | Cached read | Cache write | Output |
|---|---|---|---|---|
| Standard | $10.00 | $1.00 | $12.50 | $50.00 |
| Batch | $5.00 | $0.50 | $6.25 | $25.00 |
| Flex | $5.00 | $0.50 | $6.25 | $25.00 |
| Fast | $20.00 | $2.00 | $25.00 | $100.00 |
For Astra, Sol and Luna, Batch and Flex cost half the Standard rate; Fast costs twice the applicable rate. These ratios are specific to the verified GPT-6 rates. EU data residency requires Standard processing.
Long-context pricing
For all three models, above 272,000 total input tokens, input and cache rates double and output costs 1.5× for the full request. The table below shows Astra. At exactly the threshold, short-context rates still apply.
| Mode | Input | Cached read | Cache write | Output |
|---|---|---|---|---|
| Standard | $20.00 | $2.00 | $25.00 | $75.00 |
| Batch | $10.00 | $1.00 | $12.50 | $37.50 |
| Flex | $10.00 | $1.00 | $12.50 | $37.50 |
| Fast | $40.00 | $4.00 | $50.00 | $150.00 |
Published limits: 1,050,000 combined tokens, at most 1,050,000 input tokens and 128,000 output tokens.
Three fixed estimates
- 10,000 ordinary input + 2,000 billable output: $0.20 — ordinary input $0.10 + cached reads $0.00 + output $0.10.
- 20,000 ordinary input + 80,000 cached reads + 2,000 billable output: $0.38 — ordinary input $0.20 + cached reads $0.08 + output $0.10.
- 300,000 ordinary input + 10,000 billable output (long context): $6.75 — ordinary input $6.00 + cached reads $0.00 + output $0.75.
The cached example assumes an existing eligible cache; an earlier cache-write charge is separate. Billable output includes reasoning tokens. Tools, media, taxes and surcharges are separate.
Standard model price comparison
| Model | Input / 1M | Output / 1M | 10K input + 2K output |
|---|---|---|---|
| GPT-6 Astra | $10.00 | $50.00 | $0.20 |
| GPT-6 Sol | $2.00 | $10.00 | $0.04 |
| GPT-6 Luna | $0.10 | $0.50 | $0.002 |
| GPT-5.6 Sol | $4.00 | $20.00 | $0.08 |
| GPT-4o | $2.50 | $10.00 | $0.045 |
| GPT-4.1 | $2.00 | $8.00 | $0.036 |
Compare Astra, Sol and Luna → · Estimate cache savings →
Calculate your actual API usage cost → · Compare GPT-6 and GPT-5 for your usage
GPT 6 pricing: API bills versus subscriptions
GPT 6 pricing has two separate bills: API usage and ChatGPT subscriptions. Developers pay for tokens; ChatGPT subscribers pay for product allowances. Paying one does not fund the other. Use the access guide to identify which product applies before comparing GPT 6 pricing.
For a quick GPT 6 pricing comparison, 10,000 ordinary input tokens plus 2,000 output tokens cost $0.20 with Astra, $0.04 with Sol or $0.002 with Luna, at Standard API rates. These GPT 6 pricing examples compare identical token counts, not the measured cost of completing an identical task. The tables above and the API pricing calculator share one model registry, keeping GPT 6 pricing inputs consistent.
ChatGPT Pro plans and Codex usage
For subscribers comparing GPT 6 pricing, Pro has three monthly USD tiers as checked September 30, 2026: $100, $200 and $500. Pro $200 is open to new subscriptions again with a lower included allowance for users ineligible for the previous allowance. In the GPT 6 pricing plan comparison, Pro $500 provides the highest included usage and GPT-6 Astra Ultrafast in Work and Codex. Buying credits on Pro $100 or $200 does not unlock Ultrafast.
The GPT 6 pricing transition matters for existing Pro $200 subscribers: subscriptions active at the eligibility cutoff, or during the preceding seven days, can retain the previous allowance through October 29, 2026 while subscribed. Afterward, the allowance decreases but the price stays $200/month. For your GPT 6 pricing decision, check the account notice for eligibility; the change does not immediately apply to every existing subscriber. OpenAI’s Pro tiers guide supplies the GPT 6 pricing transition rules.
When budgeting GPT 6 pricing for agentic work, Codex and ChatGPT Work share usage with other eligible features once their pricing takes effect; OpenAI currently names Excel on Plus and Pro. GPT 6 pricing allowances are not fixed task counts: model, context, reasoning, tools and speed affect consumption. Check Usage & Billing or Codex CLI /status for remaining allowance and reset time.
Codex speed modes and purchased credits
The Codex side of GPT 6 pricing separates included usage from purchased credits and Enterprise pay-as-you-go. The Work/Codex rate card lists these multipliers:
| Speed setting | Included allowance consumption vs Standard | Purchased-credit / Enterprise pay-as-you-go rate vs Standard |
|---|---|---|
| Fast | 2.5x | 2x |
| GPT-6 Astra Ultrafast | 8x | 6x |
For GPT 6 pricing, these are billing multipliers, not speed guarantees. Do not apply them to the API calculator: GPT 6 pricing for API Fast uses a different multiplier. Included usage is consumed first, then eligible paid credits. Credits do not fund every ChatGPT feature, and Support cannot manually reset ChatGPT or Codex limits.
GPT 6 pricing in Standard Work/Codex credits per million tokens, checked September 30:
| Model | Input credits | Cached-input credits | Output credits |
|---|---|---|---|
| GPT-6 Astra | 250 | 25 | 1,250 |
| GPT-6.1 Sol | 50 | 2.5 | 250 |
| GPT-6 Sol | 50 | 5 | 250 |
| GPT-6 Luna | 2.5 | 0.25 | 12.5 |
Codex GPT 6 pricing has no separate cache-write charge; API-key usage follows API token rates instead. Within GPT 6 pricing comparisons, GPT-6.1 Sol’s cached-input credit rate is half GPT-6 Sol’s. For GPT 6 pricing in dollars, the GPT-6.1 Sol API page explains its separate token rates. This GPT 6 pricing credit table does not guarantee an included task count.
Reading GPT 6 pricing per million tokens
API GPT 6 pricing is quoted in USD per million tokens. Split total input into ordinary input, cache reads and cache writes; these categories sum to total input. A correct GPT 6 pricing estimate never counts cached tokens twice. Billable output includes reasoning tokens as well as the visible answer, so answer length alone understates GPT 6 pricing for reasoning-heavy requests.
For Astra, Sol and Luna, cached-input GPT 6 pricing is 10% of ordinary input; cache writes cost 1.25 times ordinary input. Reusing a stable prefix can therefore lower GPT 6 pricing after its first write. The prompt caching calculator estimates write cost, later reads and break-even reuse. GPT 6 pricing becomes cache-eligible at a qualifying 1,024-token prefix. Default minimum retention is 30 minutes from the latest write or read; reads refresh retention without another write charge.
GPT 6 pricing for long context and processing modes
Long-context GPT 6 pricing applies when total input exceeds 272,000 tokens. At exactly 272,000, ordinary rates apply. Above the boundary, GPT 6 pricing doubles input, cache reads and cache writes, and multiplies output by 1.5 for the entire request. Do not calculate GPT 6 pricing by charging the premium only on tokens above the threshold.
Batch and Flex GPT 6 pricing is half Standard; API Fast is twice the applicable rate. These GPT 6 pricing modes are separate from reasoning effort and reasoning.mode. More reasoning can increase output tokens. Pro reasoning retains the same base token prices but can consume more tokens. Regional GPT 6 pricing adds a 10% premium where applicable; EU data residency supports Standard processing.
Estimate monthly GPT 6 pricing
Build a monthly GPT 6 pricing estimate from a representative request. Include system instructions, retrieved documents and tool results in total input. Enter billable output and monthly requests into the calculator. Compare typical and busy months: GPT 6 pricing scales with request volume, and retries add usage even when the first answer fails validation.
At 10,000 monthly requests with the ordinary 10,000-input/2,000-output baseline, GPT 6 pricing totals $2,000 for Astra, $400 for Sol or $20 for Luna in text tokens. These GPT 6 pricing totals assume no caching, no long-context premium and Standard processing. Compare capabilities and limits in the family comparison: the lowest GPT 6 pricing unit rate may require more retries for your task.
Text-only GPT 6 pricing estimates exclude tool calls, image/audio/video usage, fine-tuning, tax and regional fees. The calculator runs locally in your browser without a paid API request.
Verify GPT 6 pricing before choosing a model
Pair GPT 6 pricing with the Astra, Sol and Luna capability pages. For integration, follow the API setup guide. GPT 6 pricing model rows retain their actual verification dates; rebuilding the site does not refresh them. Before committing a budget, compare GPT 6 pricing against the linked official rate card and your account’s allowance notice.