LLM API pricing, compared
At the usage below (input tokens: 50 million tokens; output tokens: 10 million tokens), the cheapest LLM API plan is Mistral AI at $6.00 a month (Ministral 3 (3B)); the most expensive we price is Meta Model API at $105. OpenAI API, which the most makers' products use (736), comes to $10. Set your own usage below. Price is one input: which one fits your situation matters as much.
Mistral Large 3 ($0.50 / $1.50) is still sold but superseded by Large 4. OCR and voice models are priced per page or per minute and aren't modeled here.
Standard-tier prices for models also sold by their own labs; many more open models are listed. Cached input is billed lower.
Prices from Groq's production model list (groq.com/pricing no longer shows a price table). Llama 3.1 8B and Llama 3.3 70B are listed as contact sales.
International (BytePlus) prices for the lowest input-length tier. Flex (off-peak) online inference and batch inference cost about half. Mainland prices (Volcano Engine Ark) are in yuan.
Cache-miss input prices. Cache hits cost a fraction of a cent (Pro $0.0036, Flash $0.0028 per million) and cache writes are free for now. The Batch API halves Pro and Flash prices. Mainland prices are in yuan, and a separate Token Plan subscription covers coding tools.
Base input/output prices; prompt-cache reads are billed much lower and cache writes higher. US-only inference is 1.1x, and fast mode for Opus 5.5 is 2x. Batch processing is discounted.
International prices for the lowest input-length tier. ERNIE 5.0, which adds image input and a thinking mode, costs $0.84 / $3.38.
Standard-tier prices for short context (up to 272K input tokens); long-context requests cost more (Astra $20 / $75, Sol $4 / $15, Luna $0.20 / $0.75 per million). Batch is half price; cached input is billed lower.
Cached input is billed lower (Hy4 preview $0.042, Hy3 $0.033 per million). Hy4 is still a preview release.
International (Singapore) list prices for the lowest input-length tier. Qwen3.7-Plus is marked "limited-time 20% off" on the page; the list price is shown. Batch calls are half price.
Serverless standard prices for the models also sold by their own labs. Models not listed are priced by size, from $0.10 per million tokens under 4B parameters.
International (Z.ai) prices, flat across input lengths. GLM-4.7-Flash, GLM-4.5-Flash and GLM-4.6V-Flash are free. Mainland prices (bigmodel.cn) are in yuan and tiered by length for some models.
Serverless list prices for the models also sold by their own labs (Kimi K3's time-limited promo price is left out); many more open models are listed on the page. Cached input is billed lower where shown, and the Batch API is cheaper.
A rate-limited free trial (about 1M tokens a day) and a one-time $5 credit are available. Qwen 3.8 27B is the other shared model, at $0.99 / $1.49.
Peak-hour prices with cache-miss input (peak is 01:00–04:00 and 06:00–10:00 UTC, Monday–Friday). All other hours are half price (v4-pro $0.66 / $1.98, flash $0.15 / $0.60), and cache-hit input is far cheaper, so most workloads pay less than shown.
Cache reads cost $0.06 per million. The small MiniMax-M3.1-Flash-Preview is sold only through subscription plans, not pay-as-you-go. Mainland prices (platform.minimaxi.com) are in yuan.
Paid-tier Standard text prices. Gemini 3.8 Flash is $0.75 / $3.75 through December 31, 2026 and rises to $1.50 / $7.50 on January 1, 2027. Gemini 3.1 Pro Preview costs $4 / $18 for prompts over 200K tokens. Batch and Flex are 50% off. A free tier exists but may use prompts to improve Google's products.
Prices for prompts under 200K tokens; at 200K and above input and output prices double. Web and X search tools are billed separately as tool calls.
Cache-miss input prices. Cache hits are far cheaper (K3 $0.30, K2.6 $0.16, K2.7 Code $0.19 per million). A HighSpeed variant of K2.7 Code costs twice as much. Mainland prices (platform.moonshot.cn) are in yuan.
Standard-tier prices for muse-spark-1.3 (also 1.2 and 1.1); cached input is $0.15 per million and there is no long-context premium. The contributor tier is left out because it trades your data for the lower price.
Read from each vendor's pricing page (oldest check 2026-10-07). The model is plan fee + usage beyond what's included × the vendor's rate; volume discounts, annual billing, taxes and regional prices are left out — see each tool's notes. Always confirm on the vendor's page before you commit.
Cost as you grow
Each tool's cheapest usable plan as the product grows.
At three scales
| Tool | Side projectinput tokens 5 million tokens, output tokens 1 million tokens | Growinginput tokens 50 million tokens, output tokens 10 million tokens | Scalinginput tokens 1 billion tokens, output tokens 200 million tokens |
|---|---|---|---|
| $0.60Ministral 3 (3B) | $6.00Ministral 3 (3B) | $120Ministral 3 (3B) | |
| DeepInfra | $0.63DeepSeek V4 Flash | $6.30DeepSeek V4 Flash | $126DeepSeek V4 Flash |
| $0.68GPT OSS 20B | $6.75GPT OSS 20B | $135GPT OSS 20B | |
| Doubao API | $0.90Seed 2.0 Mini | $9.00Seed 2.0 Mini | $180Seed 2.0 Mini |
| Xiaomi MiMo API | $0.98MiMo-V2.6-Flash | $9.80MiMo-V2.6-Flash | $196MiMo-V2.6-Flash |
| $1.00Claude Haiku 5.5 | $10Claude Haiku 5.5 | $200Claude Haiku 5.5 | |
| ERNIE API | $1.00ERNIE 4.5 Turbo | $10ERNIE 4.5 Turbo | $200ERNIE 4.5 Turbo |
| $1.00GPT-6 Luna | $10GPT-6 Luna | $200GPT-6 Luna | |
| Hunyuan API | $1.19Hy3 | $12Hy3 | $238Hy3 |
| Qwen API | $1.22Qwen3.8-Flash | $12Qwen3.8-Flash | $244Qwen3.8-Flash |
| $1.25GLM 5.3 Flash | $13GLM 5.3 Flash | $250GLM 5.3 Flash | |
| GLM API | $1.25GLM-5.3-Flash | $13GLM-5.3-Flash | $250GLM-5.3-Flash |
| $1.35gpt-oss-120B | $14gpt-oss-120B | $270gpt-oss-120B | |
| Cerebras Inference API | $2.50gpt-oss-120B | $25gpt-oss-120B | $500gpt-oss-120B |
| $2.70deepseek-flash | $27deepseek-flash | $540deepseek-flash | |
| MiniMax | $2.70MiniMax-M3 | $27MiniMax-M3 | $540MiniMax-M3 |
| $2.75Gemini 3.1 Flash-Lite | $28Gemini 3.1 Flash-Lite | $550Gemini 3.1 Flash-Lite | |
| $8.75Grok 4.3 | $88Grok 4.3 | $1,750Grok 4.3 | |
| Kimi API | $8.75Kimi K2.6 | $88Kimi K2.6 | $1,750Kimi K2.6 |
| Meta Model API | $11Muse Spark | $105Muse Spark | $2,100Muse Spark |
Every plan
What each LLM API plan costs, includes and caps, as modelled here.
Mistral AIPay per token, from $0.10 per million tokens on Ministral 3 (3B) to $1.50 / $7.50 per million input/output tokens on Mistral Medium 3.5.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Mistral Large 4 | None | $1.36 per million tokens | $4.18 per million tokens |
| Mistral Medium 3.5 | None | $1.5 per million tokens | $7.5 per million tokens |
| Mistral Small 4 | None | $0.15 per million tokens | $0.6 per million tokens |
| Ministral 3 (3B) | None | $0.1 per million tokens | $0.1 per million tokens |
Mistral Large 3 ($0.50 / $1.50) is still sold but superseded by Large 4. OCR and voice models are priced per page or per minute and aren't modeled here.
Checked 2026-10-19 · pricing page ↗ · Mistral AI pricing in full →DeepInfraPay per token for open-weight models, e.g. $0.09 / $0.18 per million input/output tokens on DeepSeek V4 Flash and $2.85 / $14.25 on Kimi K3.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Kimi K3 | None | $2.85 per million tokens | $14.25 per million tokens |
| Qwen3.8-Max | None | $1.65 per million tokens | $4.951 per million tokens |
| DeepSeek V4 Pro | None | $1.3 per million tokens | $2.6 per million tokens |
| DeepSeek V4 Flash | None | $0.09 per million tokens | $0.18 per million tokens |
| Nemotron 3 Ultra | None | $0.5 per million tokens | $2.2 per million tokens |
| Muse Glimmer 30B | None | $0.3 per million tokens | $1.2 per million tokens |
Standard-tier prices for models also sold by their own labs; many more open models are listed. Cached input is billed lower.
Checked 2026-10-11 · pricing page ↗ · DeepInfra pricing in full →
GroqPay per token on production models: $0.15 / $0.60 per million input/output tokens for GPT OSS 120B and $0.075 / $0.30 for GPT OSS 20B.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| GPT OSS 120B | None | $0.15 per million tokens | $0.6 per million tokens |
| GPT OSS 20B | None | $0.075 per million tokens | $0.3 per million tokens |
Prices from Groq's production model list (groq.com/pricing no longer shows a price table). Llama 3.1 8B and Llama 3.3 70B are listed as contact sales.
Checked 2026-10-11 · pricing page ↗ · Groq pricing in full →Doubao APIPay per token on BytePlus ModelArk, from $0.10 / $0.40 per million input/output tokens on Seed 2.0 Mini to $0.50 / $3 on Seed 2.0 Pro.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Seed 2.0 Pro | None | $0.5 per million tokens | $3 per million tokens |
| Seed 2.1 Turbo | None | $0.5 per million tokens | $2.5 per million tokens |
| Seed 2.0 Lite | None | $0.25 per million tokens | $2 per million tokens |
| Seed 2.0 Mini | None | $0.1 per million tokens | $0.4 per million tokens |
International (BytePlus) prices for the lowest input-length tier. Flex (off-peak) online inference and batch inference cost about half. Mainland prices (Volcano Engine Ark) are in yuan.
Checked 2026-10-19 · pricing page ↗ · Doubao API pricing in full →Xiaomi MiMo APIPay per token from a prepaid balance, from $0.14 / $0.28 per million input/output tokens on MiMo-V2.6-Flash to $0.435 / $0.87 on MiMo-V2.6-Pro.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| MiMo-V2.6-Pro | None | $0.435 per million tokens | $0.87 per million tokens |
| MiMo-V2.6-Flash | None | $0.14 per million tokens | $0.28 per million tokens |
Cache-miss input prices. Cache hits cost a fraction of a cent (Pro $0.0036, Flash $0.0028 per million) and cache writes are free for now. The Batch API halves Pro and Flash prices. Mainland prices are in yuan, and a separate Token Plan subscription covers coding tools.
Checked 2026-10-19 · pricing page ↗ · Xiaomi MiMo API pricing in full →
ClaudePay per token with no monthly fee, from $0.10 / $0.50 per million input/output tokens on Haiku 5.5 to $10 / $50 on Fable 5.1.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Claude Fable 5.1 | None | $10 per million tokens | $50 per million tokens |
| Claude Opus 5.5 | None | $4 per million tokens | $20 per million tokens |
| Claude Sonnet 5.5 | None | $2 per million tokens | $10 per million tokens |
| Claude Haiku 5.5 | None | $0.1 per million tokens | $0.5 per million tokens |
Base input/output prices; prompt-cache reads are billed much lower and cache writes higher. US-only inference is 1.1x, and fast mode for Opus 5.5 is 2x. Batch processing is discounted.
Checked 2026-10-11 · pricing page ↗ · Claude pricing in full →ERNIE APIPay per token on Qianfan, from $0.11 / $0.45 per million input/output tokens on ERNIE 4.5 Turbo to $0.56 / $2.53 on ERNIE 5.1.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| ERNIE 5.1 | None | $0.56 per million tokens | $2.53 per million tokens |
| ERNIE 4.5 Turbo | None | $0.11 per million tokens | $0.45 per million tokens |
International prices for the lowest input-length tier. ERNIE 5.0, which adds image input and a thinking mode, costs $0.84 / $3.38.
Checked 2026-10-19 · pricing page ↗ · ERNIE API pricing in full →
OpenAI APIPay per token, from $0.10 / $0.50 per million input/output tokens on GPT-6 Luna to $10 / $50 on GPT-6 Astra.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| GPT-6 Astra | None | $10 per million tokens | $50 per million tokens |
| GPT-6.1 Sol | None | $2 per million tokens | $10 per million tokens |
| GPT-6 Luna | None | $0.1 per million tokens | $0.5 per million tokens |
Standard-tier prices for short context (up to 272K input tokens); long-context requests cost more (Astra $20 / $75, Sol $4 / $15, Luna $0.20 / $0.75 per million). Batch is half price; cached input is billed lower.
Checked 2026-10-19 · pricing page ↗ · OpenAI API pricing in full →Hunyuan APIPay per token, from $0.132 / $0.528 per million input/output tokens on Hy3 to $0.834 / $2.501 on Hy4 preview.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Hy4 preview | None | $0.834 per million tokens | $2.501 per million tokens |
| Hy3 | None | $0.132 per million tokens | $0.528 per million tokens |
Cached input is billed lower (Hy4 preview $0.042, Hy3 $0.033 per million). Hy4 is still a preview release.
Checked 2026-10-19 · pricing page ↗ · Hunyuan API pricing in full →Qwen APIPay per token, from $0.15 / $0.47 per million input/output tokens on Qwen3.8-Flash to $2 / $6 on Qwen3.8-Max (international region).
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Qwen3.8-Max | None | $2 per million tokens | $6 per million tokens |
| Qwen3.7-Plus | None | $0.4 per million tokens | $1.6 per million tokens |
| Qwen3.8-Flash | None | $0.15 per million tokens | $0.47 per million tokens |
International (Singapore) list prices for the lowest input-length tier. Qwen3.7-Plus is marked "limited-time 20% off" on the page; the list price is shown. Batch calls are half price.
Checked 2026-10-19 · pricing page ↗ · Qwen API pricing in full →
Fireworks AIPay per token for other labs' open-weight models, e.g. $0.15 / $0.60 per million input/output tokens on gpt-oss-120B and $3 / $15 on Kimi K3.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Kimi K3 | None | $3 per million tokens | $15 per million tokens |
| Qwen 3.8 Max | None | $2 per million tokens | $6 per million tokens |
| GLM 5.3 | None | $1.4 per million tokens | $4.4 per million tokens |
| MiniMax M3 | None | $0.3 per million tokens | $1.2 per million tokens |
| DeepSeek V4.1 Flash | None | $0.3 per million tokens | $1.2 per million tokens |
| GLM 5.3 Flash | None | $0.15 per million tokens | $0.5 per million tokens |
| gpt-oss-120B | None | $0.15 per million tokens | $0.6 per million tokens |
Serverless standard prices for the models also sold by their own labs. Models not listed are priced by size, from $0.10 per million tokens under 4B parameters.
Checked 2026-10-19 · pricing page ↗ · Fireworks AI pricing in full →GLM APIPay per token, from $0.15 / $0.50 per million input/output tokens on GLM-5.3-Flash to $1.40 / $4.40 on GLM-5.3, with some older Flash models free.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| GLM-5.3 | None | $1.4 per million tokens | $4.4 per million tokens |
| GLM-5.3-FlashX | None | $0.37 per million tokens | $1.25 per million tokens |
| GLM-5.3-Flash | None | $0.15 per million tokens | $0.5 per million tokens |
International (Z.ai) prices, flat across input lengths. GLM-4.7-Flash, GLM-4.5-Flash and GLM-4.6V-Flash are free. Mainland prices (bigmodel.cn) are in yuan and tiered by length for some models.
Checked 2026-10-19 · pricing page ↗ · GLM API pricing in full →
Together AIPay per token for other labs' open-weight models, e.g. $0.15 / $0.60 per million input/output tokens on gpt-oss-120B and $1.40 / $4.40 on GLM-5.3.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Kimi K3 | None | $3 per million tokens | $15 per million tokens |
| GLM-5.3 | None | $1.4 per million tokens | $4.4 per million tokens |
| DeepSeek V4 Pro | None | $1.32 per million tokens | $3.96 per million tokens |
| DeepSeek V4.1 Flash | None | $0.3 per million tokens | $1.2 per million tokens |
| gpt-oss-120B | None | $0.15 per million tokens | $0.6 per million tokens |
| Muse Glimmer 30B | None | $0.35 per million tokens | $1.5 per million tokens |
Serverless list prices for the models also sold by their own labs (Kimi K3's time-limited promo price is left out); many more open models are listed on the page. Cached input is billed lower where shown, and the Batch API is cheaper.
Checked 2026-10-11 · pricing page ↗ · Together AI pricing in full →Cerebras Inference APIPay per token on very fast inference: $0.35 / $0.75 per million input/output tokens on gpt-oss-120B.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| gpt-oss-120B | None | $0.35 per million tokens | $0.75 per million tokens |
A rate-limited free trial (about 1M tokens a day) and a one-time $5 credit are available. Qwen 3.8 27B is the other shared model, at $0.99 / $1.49.
Checked 2026-10-19 · pricing page ↗ · Cerebras Inference API pricing in full →
DeepSeekPay per token from a prepaid balance, with off-peak hours at half price and cache-hit input far cheaper.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| deepseek-v4-pro | None | $1.32 per million tokens | $3.96 per million tokens |
| deepseek-flash | None | $0.3 per million tokens | $1.2 per million tokens |
Peak-hour prices with cache-miss input (peak is 01:00–04:00 and 06:00–10:00 UTC, Monday–Friday). All other hours are half price (v4-pro $0.66 / $1.98, flash $0.15 / $0.60), and cache-hit input is far cheaper, so most workloads pay less than shown.
Checked 2026-10-11 · pricing page ↗ · DeepSeek pricing in full →MiniMaxPay per token: $0.30 / $1.20 per million input/output tokens on MiniMax-M3 for requests up to 512K input tokens.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| MiniMax-M3 | None | $0.3 per million tokens | $1.2 per million tokens |
Cache reads cost $0.06 per million. The small MiniMax-M3.1-Flash-Preview is sold only through subscription plans, not pay-as-you-go. Mainland prices (platform.minimaxi.com) are in yuan.
Checked 2026-10-07 · pricing page ↗ · MiniMax pricing in full →
Gemini APIPay per token on the paid tier, from $0.25 / $1.50 per million input/output tokens on Gemini 3.1 Flash-Lite to $2 / $12 on 3.1 Pro Preview; a free tier covers some models.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Gemini 3.1 Pro Preview | None | $2 per million tokens | $12 per million tokens |
| Gemini 3.8 Flash | None | $0.75 per million tokens | $3.75 per million tokens |
| Gemini 3.5 Flash-Lite | None | $0.3 per million tokens | $2.5 per million tokens |
| Gemini 3.1 Flash-Lite | None | $0.25 per million tokens | $1.5 per million tokens |
Paid-tier Standard text prices. Gemini 3.8 Flash is $0.75 / $3.75 through December 31, 2026 and rises to $1.50 / $7.50 on January 1, 2027. Gemini 3.1 Pro Preview costs $4 / $18 for prompts over 200K tokens. Batch and Flex are 50% off. A free tier exists but may use prompts to improve Google's products.
Checked 2026-10-11 · pricing page ↗ · Gemini API pricing in full →
Grok APIPay per token: $2 / $6 per million input/output tokens on Grok 4.7 and $1.25 / $2.50 on Grok 4.3.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Grok 4.7 | None | $2 per million tokens | $6 per million tokens |
| Grok 4.3 | None | $1.25 per million tokens | $2.5 per million tokens |
Prices for prompts under 200K tokens; at 200K and above input and output prices double. Web and X search tools are billed separately as tool calls.
Checked 2026-10-19 · pricing page ↗ · Grok API pricing in full →Kimi APIPay per token from a prepaid balance, from $0.95 / $4 per million input/output tokens on Kimi K2.6 to $3 / $15 on Kimi K3.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Kimi K3 | None | $3 per million tokens | $15 per million tokens |
| Kimi K2.6 | None | $0.95 per million tokens | $4 per million tokens |
| Kimi K2.7 Code | None | $0.95 per million tokens | $4 per million tokens |
Cache-miss input prices. Cache hits are far cheaper (K3 $0.30, K2.6 $0.16, K2.7 Code $0.19 per million). A HighSpeed variant of K2.7 Code costs twice as much. Mainland prices (platform.moonshot.cn) are in yuan.
Checked 2026-10-19 · pricing page ↗ · Kimi API pricing in full →Meta Model APIPay per token, no minimums: $1.25 / $4.25 per million input/output tokens on Muse Spark, or far less on a tier that lets Meta train on your data.
| Plan | Monthly fee | Input tokens per month | Output tokens per month |
|---|---|---|---|
| Muse Spark | None | $1.25 per million tokens | $4.25 per million tokens |
Standard-tier prices for muse-spark-1.3 (also 1.2 and 1.1); cached input is $0.15 per million and there is no long-context premium. The contributor tier is left out because it trades your data for the lower price.
Checked 2026-10-11 · pricing page ↗ · Meta Model API pricing in full →