LLM API pricing comparison (October 2026)
LLM API list prices in October 2026 run from $0.10 to $10 per million input tokens and from $0.28 to $50 per million output tokens on the model makers' own pricing pages, and the everyday coding tier sits at $2 input: Claude Sonnet 5.5 at $2 / $10 and GPT-5.6 Terra at $2 / $12.
| Model | Maker (pricing page) | Input | Cached input | Output | Long prompts, promotions |
|---|---|---|---|---|---|
| GPT-6 Astraopenai/gpt-6-astra | OpenAI | $10 | $1 | $50 | Over 272K: $20 / $75 |
| GPT-5.6 Solopenai/gpt-5.6-sol | OpenAI | $4 | $0.40 | $20 | Over 272K: $8 / $30OpenAI calls this promotional pricing, available at least through Nov 21, 2026 |
| GPT-5.6 Terraopenai/gpt-5.6-terra | OpenAI | $2 | $0.20 | $12 | Over 272K: $4 / $18 |
| GPT-5.6 Lunaopenai/gpt-5.6-luna | OpenAI | $0.20 | $0.02 | $1.20 | Over 272K: $0.40 / $1.80 |
| GPT-6 Lunaopenai/gpt-6-luna | OpenAI | $0.10 | $0.01 | $0.50 | Over 272K: $0.20 / $0.75 |
| Claude Fable 5.1anthropic/claude-fable-5.1 | Anthropic | $10 | $0.25 | $50 | |
| Claude Fable 5anthropic/claude-fable-5 | Anthropic | $10 | $1 | $50 | |
| Claude Opus 5anthropic/claude-opus-5 | Anthropic | $5 | $0.50 | $25 | |
| Claude Opus 5.5anthropic/claude-opus-5.5 | Anthropic | $4 | $0.20 | $20 | |
| Claude Sonnet 5.5anthropic/claude-sonnet-5.5 | Anthropic | $2 | $0.10 | $10 | |
| Claude Sonnet 5anthropic/claude-sonnet-5 | Anthropic | $2 | $0.20 | $10 | |
| Claude Haiku 5.5anthropic/claude-haiku-5.5 | Anthropic | $0.10 | $0.01 | $0.50 | Over 100K: $0.50 / $2.50 |
| Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview | $2 | $0.20 | $12 | Over 200K: $4 / $18 | |
| Gemini 3.8 Flashgoogle/gemini-3.8-flash | $1.50 | $0.15 | $7.50 | Half price ($0.75 / $3.75) through Dec 31, 2026 | |
| Gemini 2.5 Progoogle/gemini-2.5-pro | $1.25 | $0.125 | $10 | Over 200K: $2.50 / $15 | |
| Gemini 3.5 Flash-Litegoogle/gemini-3.5-flash-lite | $0.30 | $0.03 | $2.50 | ||
| Gemini 2.5 Flashgoogle/gemini-2.5-flash | $0.30 | $0.03 | $2.50 | ||
| Kimi K3moonshotai/kimi-k3 | Moonshot | $3 | $0.30 | $15 | |
| Qwen 3.8 Maxqwen/qwen3.8-max-0902 | Alibaba | $2 | Not listed | $6 | |
| Qwen 3.8 27Bqwen/qwen3.8-27b | Alibaba | $0.50 | Not listed | $3 | |
| Qwen 3.7 Plusqwen/qwen3.7-plus | Alibaba | $0.40 | Not listed | $1.60 | Over 256K: $1.20 / $4.8020% off for a limited time ($0.32 / $1.28) |
| Qwen 3.8 Flashqwen/qwen3.8-flash | Alibaba | $0.15 | Not listed | $0.47 | |
| GLM 5.3z-ai/glm-5.3 | Z.ai | $1.40 | $0.26 | $4.40 | |
| GLM 5.2z-ai/glm-5.2 | Z.ai | $1.40 | $0.26 | $4.40 | |
| GLM 5.3 Flashz-ai/glm-5.3-flash | Z.ai | $0.15 | $0.03 | $0.50 | |
| DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813 | DeepSeek | $1.32 | $0.044 | $3.96 | Peak hours; half price off-peak |
| DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash | DeepSeek | $0.30 | $0.006 | $1.20 | Peak hours; half price off-peak |
| Mistral Large 4mistralai/mistral-large-4-0 | Mistral | $1.36 | $0.14 | $4.18 | On sale at half price ($0.68 / $2.09), no end date shown |
| Grok 4.3x-ai/grok-4.3 | xAI | $1.25 | $0.20 | $2.50 | 200K and over: $2.50 / $5 |
| MiniMax M3minimax/minimax-m3 | MiniMax | $0.30 | $0.06 | $1.20 | Over 512K: $0.60 / $2.40 |
| MiMo V2.6 Proxiaomi/mimo-v2.6-pro | Xiaomi | $0.435 | $0.0036 | $0.87 | |
| MiMo V2.6 Flashxiaomi/mimo-v2.6-flash | Xiaomi | $0.14 | $0.0028 | $0.28 |
What moves the bill besides the list price
- Cached input. Coding agents resend the same context on every step, so the cache-read price often matters more than the input price: Claude Sonnet 5.5 charges $0.10 per million for cache reads against $2 for fresh input.
- Long prompts. OpenAI charges a higher rate for the whole request once the prompt passes 272K tokens, Google above 200K on its Pro models, Anthropic above 100K on Haiku 5.5 only, xAI from 200K and MiniMax above 512K.
- Promotions. GPT-5.6 Sol, Gemini 3.8 Flash, Mistral Large 4 and Qwen 3.7 Plus list a discounted price today, and three of them name an end date or call it limited.
- Time of day. DeepSeek charges half outside its peak hours (01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays).
New models not in this table
Three new models are not served on Standard Compute yet, so they are not in the table. GPT-6 Sol (September 22) and GPT-6.1 Sol (September 29) both list at $2 input and $10 output on OpenAI's API; the closest served OpenAI models are GPT-6 Astra and GPT-5.6 Sol. Gemini 4 Argon, released on October 13, lists at $4 / $20 with half price through November 30, 2026; the closest served Google model is Gemini 3.1 Pro Preview.
How we checked
Every price was read on October 14, 2026 on the model maker's own pricing page or model documentation, linked from each maker name, never on OpenRouter, a reseller or a third-party comparison. Where the maker shows a discount, the table gives its standard rate and the note gives the discount; GPT-5.6 Sol is the exception, because OpenAI's page lists only its promotional rate. Model IDs are the exact IDs served at https://api.stdcmpt.com/v1.
To correct a row, email contact@standardcompute.com with a link to the maker's page and we will update the row and its date.
Where Standard Compute fits
Per-token prices suit light, bursty use; an agent that runs most of the day spends at the rate of whichever model it calls. Standard Compute serves every model in this table through one OpenAI- and Anthropic-compatible key for a fixed monthly price, such as Standard at $89 with a $95 compute budget or Pro at $249 with $269. Each request draws from the budget at that model's per-token rate, smart routing sends routine steps to the efficient models so the budget covers more work, and at the budget requests stop with HTTP 402. Nothing is billed on top unless you turn on Pay as you go.
FAQ
What is the cheapest LLM API in October 2026?
Among models that handle coding-agent work, GPT-6 Luna and Claude Haiku 5.5 list at $0.10 input and $0.50 output per million tokens, MiMo V2.6 Flash at $0.14 / $0.28, and GLM 5.3 Flash and Qwen 3.8 Flash at $0.15 input. DeepSeek V4.1 Flash is $0.30 / $1.20 at peak and half that off-peak, with cache reads at $0.006.
How much does GPT-5.6 Sol cost per million tokens?
$4 input, $0.40 cached input and $20 output, which OpenAI calls promotional pricing available at least through November 21, 2026. Prompts over 272K tokens cost $8 / $30. GPT-5.6 Terra is $2 / $12 and GPT-5.6 Luna $0.20 / $1.20.
Is Claude Opus 5.5 cheaper than Claude Opus 5?
Yes. Opus 5.5 lists at $4 input and $20 output per million tokens with cache reads at $0.20; Opus 5 is $5 / $25 with cache reads at $0.50. Neither has a long-context surcharge.
Which LLM APIs charge more for long prompts?
OpenAI's GPT-6 and GPT-5.6 models above 272K tokens, Gemini 3.1 Pro and 2.5 Pro above 200K, Claude Haiku 5.5 above 100K, Grok 4.3 from 200K and MiniMax M3 above 512K. The higher rate applies to the whole request. Claude Opus, Sonnet and Fable, GLM, Kimi K3 and DeepSeek list no long-context surcharge.
Is there a flat monthly alternative to per-token LLM API pricing?
Yes. Standard Compute serves every model in this table behind one OpenAI- and Anthropic-compatible key for a fixed monthly price from $19, with a fixed compute budget: Standard is $89 with $95 of compute. Requests stop with HTTP 402 at the budget, and there is no overage bill unless you turn on Pay as you go.