API or subscription?
Enter your monthly tokens to see the API bill for each model next to the price of the subscriptions we track. Official per-token prices, standard tier.
Last verified 2026-10-19
Official API prices per 1M tokens
| Provider | Model | Input | Cached input | Output | Verified |
|---|---|---|---|---|---|
| Anthropic | Claude Opus 5.5 | $4 | $0.2 | $20 | 2026-10-19 source |
| Anthropic | Claude Sonnet 5.5 | $2 | $0.1 | $10 | 2026-10-19 source |
| Anthropic | Claude Haiku 5.5 Prompts up to 100K tokens; $0.50/$2.50 above that | $0.1 | $0.01 | $0.5 | 2026-10-19 source |
| Anthropic | Claude Opus 5 | $5 | $0.5 | $25 | 2026-10-19 source |
| Anthropic | Claude Sonnet 5 $2/$10 is now the standard price | $2 | $0.2 | $10 | 2026-10-19 source |
| Anthropic | Claude Opus 4.8 | $5 | $0.5 | $25 | 2026-10-19 source |
| Anthropic | Claude Sonnet 4.6 | $3 | $0.3 | $15 | 2026-10-19 source |
| Anthropic | Claude Haiku 4.5 | $1 | $0.1 | $5 | 2026-10-19 source |
| OpenAI | gpt-6.1-sol Up to 272K input | $2 | $0.1 | $10 | 2026-10-19 source |
| OpenAI | gpt-6-astra Up to 272K input | $10 | $1 | $50 | 2026-10-19 source |
| OpenAI | gpt-6-luna Up to 272K input | $0.1 | $0.01 | $0.5 | 2026-10-19 source |
| OpenAI | gpt-5.6-sol Promotional price at least through Nov 21, 2026 | $4 | $0.4 | $20 | 2026-10-19 source |
| OpenAI | gpt-5.3-codex | $1.75 | $0.175 | $14 | 2026-10-19 source |
| OpenAI | gpt-5 | $1.25 | $0.125 | $10 | 2026-10-19 source |
| OpenAI | gpt-5-mini | $0.25 | $0.025 | $2 | 2026-10-19 source |
| gemini-4-argon Through Nov 30, 2026; $4/$20/$0.20 from Dec 1, 2026. Not on the free tier | $2 | $0.1 | $10 | 2026-10-19 source | |
| gemini-3.8-flash Through Dec 31, 2026; $1.50/$7.50/$0.15 from Jan 1, 2027 | $0.75 | $0.075 | $3.75 | 2026-10-19 source | |
| gemini-3.1-pro-preview Prompts up to 200K tokens | $2 | $0.2 | $12 | 2026-10-19 source | |
| gemini-2.5-pro Prompts up to 200K tokens | $1.25 | $0.125 | $10 | 2026-10-19 source | |
| gemini-2.5-flash Text/image/video input | $0.3 | $0.03 | $2.50 | 2026-10-19 source | |
| DeepSeek | deepseek-flash (V4.1 Flash) Peak rate; off-peak is half | $0.3 | $0.006 | $1.20 | 2026-10-19 source |
| DeepSeek | deepseek-v4-pro Peak rate; off-peak is half | $1.32 | $0.044 | $3.96 | 2026-10-19 source |
| Mistral | mistral-large-4 (Mistral Large 4) Open weights (modified MIT; model-as-a-service businesses need a separate license). Price from Mistral's Oct 29, 2026 launch post. | $0.7 | $0.07 | $2.10 | 2026-10-30 source |
Assumptions
- Standard tier, short-context prices in USD per 1M tokens. Batch, flex and fast modes, long-context premiums, cache writes and data-residency uplifts are not included.
- Claude 4.7 and later models use a newer tokenizer that produces about 30% more tokens for the same text (Anthropic pricing docs).
- OpenAI's Codex page says API token prices are separate from subscription usage and shouldn't be used to estimate included tasks.
- DeepSeek prices shown are peak-hour rates; off-peak is half price.
- Anthropic charges cache writes (1.25x input for 5-minute cache) on top; this estimate doesn't include them.
- Promotional prices end on the dates noted in the table, after which costs rise.
- Not included: Qwen coder models (Alibaba Cloud Model Studio). The official pricing page didn't load for us, so prices aren't listed.
FAQ
Is the API cheaper than a subscription for coding?
It depends on how many tokens you use. Use the calculator: enter your monthly input and output tokens to see the API bill for each model next to the subscription price. Subscriptions don't publish token allowances, so we can't say how many tokens a plan really includes.
What do the API prices include?
Standard tier, short-context prices in USD per 1M tokens. Batch, flex and fast modes, long-context premiums, cache writes and data-residency uplifts are not included.
Why do my tokens differ between Claude and other models?
Anthropic says Claude 4.7 and later use a newer tokenizer that produces about 30% more tokens for the same text. The same work can count as different token totals on different models.
Which subscriptions are priced in dollars of usage?
GitHub Copilot paid plans include monthly AI credits (1 credit = $0.01): $15 on Pro and $70 on Pro+. Warp Build includes $20 of agent usage at API rates. Claude Max includes $100 to $200/mo of Claude API credits, but they can't be used in Claude Code (Anthropic says they don't apply to interactive Claude Code or extra usage).
Pick a subscription by budget · Usage limits explained · Free API tiers
Never get surprised by a price hike
The free weekly Quotadex changelog: price changes, new limits, and which plan is the best deal right now.
1 email a week. No spam. Unsubscribe anytime.