OpenAI's API lineup is now led by the GPT-6 family (Astra, Sol and Luna), with GPT-5.6, GPT-5.5 pro and GPT-5.4 models still on sale beside it. The rates below are recalculated from our dataset whenever OpenAI changes them. ChatGPT subscriptions are a separate product with separate billing.
OpenAI's cheapest current model is GPT-6 Luna at $0.10 input / $0.50 output per 1M tokens. The GPT-6.1 Sol workhorse costs $2.00 / $10.00 and the flagship GPT-6 Astra $10.00 / $50.00. Prompts over 272K tokens are billed at a higher long-context rate, and the Batch and Flex tiers cost half the standard price.
| Model | Tier | Input / 1M | Output / 1M | Cache hit / 1M | Context | Notes |
|---|---|---|---|---|---|---|
| GPT-6 Astra | Flagship | $10.00 | $50.00 | $1.00 | 1M | New top flagship (Sep 3, 2026); rolling out to Plus/Pro/Business/Enterprise; >272K tokens: $20/$75, cache $2 |
| GPT-6.1 Sol | Mid-tier | $2.00 | $10.00 | $0.10 | 1M | New (Sep 29, 2026 DevDay), below Astra in the GPT-6 family; replaces GPT-6 Sol at the same input/output price with a 50% cheaper cache-read rate; >272K tokens: $4/$15, cache $0.20. ChatGPT Work/Codex on Plus/Pro/Business/Enterprise/Edu, plus API |
| GPT-6.1 Terra | Mid-tier | $1.00 | $5.00 | $0.05 | 1M | New mid-tier GPT-6.1 model (Oct 26, 2026), 50% cheaper than both GPT-5.6 Terra and GPT-6.1 Sol; computer-use, visual reasoning & coding; cache write $1.25 (30-min) or $2.50 (12-hr); >272K tokens: $2/$7.50, cache $0.10 |
| GPT-6 Luna | Budget | $0.10 | $0.50 | $0.01 | 1M | New (Sep 29, 2026 DevDay), fastest & cheapest GPT-6-family tier; >272K tokens: $0.20/$0.75, cache $0.02 |
| GPT-5.5 pro | Flagship | $30.00 | $180.00 | — | 1M | Higher-compute GPT-5.5 for hard problems (slowest); no cached-input discount; >272K tokens: $60/$270 |
| GPT-5.6 Sol | Flagship | $4.00 | $20.00 | $0.40 | 1M | Cut 20% Aug 21 (promo thru Nov 21); >272K tokens: $8/$30, cache $0.80 |
| GPT-5.6 Terra | Mid-tier | $2.00 | $12.00 | $0.20 | 1M | Balanced 5.6; cut 20% Jul 30; >272K tokens: $4/$18, cache $0.40 |
| GPT-5.6 Luna | Budget | $0.20 | $1.20 | $0.02 | 1M | Fast & cheapest 5.6 tier; cut 80% Jul 30; >272K tokens: $0.40/$1.80, cache $0.04 |
| GPT-5.4 mini | Budget | $0.75 | $4.50 | $0.075 | 400K | Cheaper, well-defined tasks |
| GPT-5.4 nano | Budget | $0.20 | $1.25 | $0.02 | 400K | Fastest, cheapest |
| Model | Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens |
|---|---|---|---|
| GPT-6 Luna | $4.00 | $3.40 | $0.15 |
| GPT-5.6 Luna | $8.80 | $6.96 | $0.33 |
| GPT-5.4 nano | $9.00 | $7.00 | $0.34 |
| GPT-5.4 mini | $33.00 | $26.10 | $1.25 |
| GPT-6.1 Terra | $40.00 | $34.00 | $1.33 |
| GPT-6.1 Sol | $80.00 | $68.00 | $2.66 |
| GPT-5.6 Terra | $88.00 | $69.60 | $3.32 |
| GPT-5.6 Sol | $160 | $136 | $6.04 |
| GPT-6 Astra | $400 | $340 | $15.10 |
| GPT-5.5 pro | $1,320 | $1,044 | $147 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
| Model | Type | Price | Notes |
|---|---|---|---|
| GPT-Image-2.5 Flare | Image | ≈$0.064 / image | New flagship (Sep 8, 2026), default fast variant; same per-token rates as GPT-Image-2 ($8 image in / $30 image out per 1M) with higher quality & lower latency |
| GPT-Image-2.5 Sunburst | Image | ≈$0.064 / image | Premium editing/control variant (Sep 8, 2026); shares Flare's published per-token price |
| GPT-Image-2 | Image | ≈$0.053 / image | Superseded by GPT-Image-2.5 Flare/Sunburst (Sep 8, 2026) as OpenAI's flagship, but still offered; photoreal edits; medium quality at 1024×1024 shown (low ≈$0.006, high ≈$0.211) |
| GPT-Image-1.5 | Image | ≈$0.034 / image | Being retired Dec 1, 2026 (along with GPT-Image-1-mini) — migrate to GPT-Image-2.5 |
| GPT-Image-1-mini | Image | ≈$0.011 / image | Cheap OpenAI tier; being retired Dec 1, 2026 along with GPT-Image-1.5 |
| GPT-Realtime-2.1 | Voice | $32 in / $64 out per 1M audio tok | Speech-to-speech agents with tool use |
| GPT-Live-1 | Voice | $0.05 / min (voice layer) | New (Sep 10, 2026), full-duplex real-time voice; billed per second, backend reasoning model billed separately |
| TTS-1 HD | Voice | $30 / 1M chars | HD text-to-speech; shuts down Jan 6, 2027 — OpenAI's replacement is gpt-realtime-2.1-mini |
| TTS-1 | Voice | $15 / 1M chars | Standard text-to-speech; shuts down Jan 6, 2027 — OpenAI's replacement is gpt-realtime-2.1-mini |
| Whisper | Voice | $0.006 / min | Transcription (speech-to-text); shuts down Feb 26, 2027 — replaced by gpt-transcribe ($0.0045/min) or gpt-live-transcribe |
| GPT-Transcribe | Voice | $0.0045 / min | Cheapest OpenAI transcription; recommended replacement for Whisper |
| Model | Provider | Tier | Input / 1M | Output / 1M | Context |
|---|---|---|---|---|---|
| GPT-5.6 Sol | OpenAI | Flagship | $4.00 | $20.00 | 1M |
| Grok 4.7 | SpaceXAI | Flagship | $2.00 | $6.00 | 500K |
| Qwen3.8-Max | Alibaba | Flagship | $2.00 | $6.00 | 1M |
| Gemini 4 Argon | Flagship | $2.00 | $10.00 | 1M | |
| Kimi K3.1 Pro | Moonshot AI | Flagship | $2.00 | $10.00 | 1M |
| GPT-6.1 Terra | OpenAI | Mid-tier | $1.00 | $5.00 | 1M |
| Mistral Large 3 | Mistral | Mid-tier | $0.50 | $1.50 | 256K |
| DeepSeek V4-Pro | DeepSeek | Mid-tier | $0.66 | $1.98 | 1M |
| Mistral Large 4 | Mistral | Mid-tier | $0.70 | $2.10 | 1M |
| Grok 4.3 | SpaceXAI | Mid-tier | $1.25 | $2.50 | 1M |
| GPT-6 Luna | OpenAI | Budget | $0.10 | $0.50 | 1M |
| Ministral 3 14B | Mistral | Budget | $0.20 | $0.20 | 256K |
| Gemini 2.5 Flash-Lite | Budget | $0.10 | $0.40 | 1M | |
| Claude Haiku 5.5 | Anthropic | Budget | $0.10 | $0.50 | 1M |
| Mistral Small 4 | Mistral | Budget | $0.15 | $0.60 | 256K |
OpenAI's cheapest model per tier (bold) against the four cheapest same-tier models from other providers, by output price.
The GPT-6 family is a three-step ladder with the same 1M context on every rung, so the choice is purely about price and capability. GPT-6 Luna ($0.50 output) is built for high-volume classification, extraction and short replies. GPT-6.1 Sol ($10.00 output) is the sensible default for chat and agents. GPT-6 Astra ($50.00 output) is worth paying for only on the steps that fail on Sol. A common pattern is to let Sol plan and call tools, and escalate to Astra only for a final review or a failed attempt. That keeps most tokens on the cheaper rate.
On the GPT-6 and GPT-5.6 models, a request whose input goes past 272K tokens is billed at long-context rates: double the input price and 1.5× the output price (GPT-6 Astra, for example, goes to $20/$75 per 1M). The 1M-token window is there when you need it, but retrieval that trims a 400K-token prompt to 200K cuts that request's input cost by three-quarters: half from sending fewer tokens, and half again from dropping back to the standard rate. The output of that request gets cheaper too. GPT-5.4 mini and nano have a smaller 400K window and no long-context tier.
OpenAI's pricing page lists Batch and Flex rates at exactly half the standard input, cached-input and output prices for the GPT-6 and GPT-5.6 models. Nightly evaluations, bulk summarisation and backfills belong there. The opposite lever is Fast mode (Priority processing was renamed Fast mode on July 30, 2026), which costs more than standard. An even faster Ultrafast tier exists only for GPT-6 Astra and GPT-6.1 Sol. The cost tables on this page use standard rates.
Mostly no. GPT-6.1 Sol charges $10.00 output against $12.00 for GPT-5.6 Terra, and GPT-6 Luna undercuts both GPT-5.6 Luna and GPT-5.4 nano on input and output. GPT-5.6 Sol is on a temporary 20% cut that our data shows running only until Nov 21, so don't build a budget on it. GPT-5.5 pro, at $180.00 output with no cache discount, only makes sense where you have tested that nothing in the GPT-6 family gets the job done.
Not for text models. OpenAI's API pricing page doesn't list a free tier for GPT models; the only model priced as free there is the moderation model. The free ChatGPT plan is a consumer app and doesn't include API access.
No. OpenAI bills ChatGPT and the API platform separately, each with its own billing settings. ChatGPT Plus ($20/mo) covers the ChatGPT app; API usage is charged per token from your API account.
On list price they're close at the low and middle tiers: GPT-6 Luna ($0.10 / $0.50) against Claude Haiku 5.5 ($0.10 / $0.50), and GPT-6.1 Sol ($2.00 / $10.00) against Claude Sonnet 5.5 ($2.00 / $10.00). At the top, Claude Opus 5.5 ($20.00 output) is cheaper than GPT-6 Astra ($50.00). Two things change the real bill. Anthropic says the tokenizer in Claude 4.7 and later models produces about 30% more tokens for the same text than its previous one, and token counts differ between vendors anyway, so compare cost per task on your own prompts. And OpenAI raises rates above 272K input tokens, while Sonnet 5.5 and Opus 5.5 bill their full 1M window at the standard rate.
A request is billed as input tokens × $10.00 per 1M plus output tokens × $50.00 per 1M, with repeated prompt prefixes billed at the cached rate of $1.00. Output dominates on long answers. A reply of 2,000 tokens costs as much as 10,000 tokens of fresh input, so capping output length is the quickest saving.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.