Kimi K2 Thinking API pricing
Updated
Among the 6 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest blended standard price for Kimi K2 Thinking is GMI Cloud at USD 0.80 per 1M input tokens and USD 1.20 per 1M output tokens. The median across 6 standard prices is USD 0.60 / USD 2.50. Each price is read from the provider's own pricing page.
Does Moonshot AI sell Kimi K2 Thinking directly?
Moonshot AI publishes no price for Kimi K2 Thinking on its own API that we could find, so this page compares hosts only.
Kimi K2 Thinking price per provider
USD per 1M tokens. Standard prices are ranked on a blend of 3 input tokens to 1 output token, cheapest first; the dearest standard price is 19% above the cheapest.
| # | Provider | Input | Cached input | Output | Blended 3:1 | Context | Precision | Source |
|---|---|---|---|---|---|---|---|---|
| 1 | GMI Cloud cheapestmoonshotai/Kimi-K2-ThinkingNotes and source quoteNo cache-read price published. “Price (per 1M tokens) Input $0.80 Output $1.20” | USD 0.80 | not published | USD 1.20 | USD 0.90 | 262,144 | INT4 | GMI Cloud page checked 2026-11-15 re-checked 2026-11-16 |
| 2 | Amazon BedrockNotes and source quoteTable headed "Regions: US East (N. Virginia), US East (Ohio) and US West (Oregon)". “Kimi K2 Thinking $ 0.60 $ 2.50” | USD 0.60 | not published | USD 2.50 | USD 1.075 | not stated | not stated | Amazon Bedrock page checked 2026-11-14 re-checked 2026-11-16 |
| 3 | Featherlessmoonshotai/Kimi-K2-ThinkingNotes and source quoteDeveloper plan (per-token billing, from 50 USD a month in credits). No cached price published. “PRICING: INPUT $0.6 / OUTPUT $2.5” | USD 0.60 | not published | USD 2.50 | USD 1.075 | 32,768 | FP8 | Featherless page checked 2026-11-15 re-checked 2026-11-16 |
| 4 | Google Vertex AIkimi-k2-thinkingNotes and source quote“Kimi-K2-Thinking Input $0.60 … Output $2.50 … Cache Hit $0.06” | USD 0.60 | USD 0.06 | USD 2.50 | USD 1.075 | not stated | not stated | Google Vertex AI page checked 2026-11-14 re-checked 2026-11-16 |
| 5 | Microsoft AzureNotes and source quoteFoundry Models serverless (pay-as-you-go), Global deployment. “Kimi K2 Thinking Global $0.60 N/A $2.50” | USD 0.60 | not published | USD 2.50 | USD 1.075 | not stated | not stated | Microsoft Azure page checked 2026-11-14 re-checked 2026-11-16 |
| 6 | Novita AImoonshotai/kimi-k2-thinkingNotes and source quote“Kimi K2 Thinking 256K $0.6 /Mt · Cache Read $0.15 /Mt $2.5 /Mt” | USD 0.60 | USD 0.15 | USD 2.50 | USD 1.075 | 262,144 | not stated | Novita AI page checked 2026-11-14 re-checked 2026-11-16 |
- Ranked: standard-tier prices only, on (3 × input + output) ÷ 4. "Cached input: not published" means the provider's page shows no cache price; budget cached tokens at the input price.
- Automated re-check 2026-11-16: a weekly job re-fetches each source page and looks for the recorded quote or prices. It is not a human review. "Re-checked" means it still found them. Rows with no re-check note show only the date the price was read.
- Precision is shown only where the provider states it; "not stated" is unknown, not full precision.
Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).
What does Kimi K2 Thinking cost for a real workload?
Monthly cost over 30 days at each standard price, using the presets of the LLM API cost calculator. Providers without a cached price bill cached input at the input price.
- Support chatbot: 2,000 requests a day, 1,500 input tokens (60% cached), 300 output.
- RAG over documents: 1,000 requests a day, 8,000 input tokens (30% cached), 500 output.
- Coding agent (one developer): 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning.
| Provider | Support chatbot | RAG over documents | Coding agent (one developer) |
|---|---|---|---|
| GMI Cloud | USD 93.60 | USD 210.00 | USD 1,548 |
| Amazon Bedrock | USD 99.00 | USD 181.50 | USD 1,305 |
| Featherless | USD 99.00 | USD 181.50 | exceeds context |
| Google Vertex AI | USD 69.84 | USD 142.62 | USD 478.80 |
| Microsoft Azure | USD 99.00 | USD 181.50 | USD 1,305 |
| Novita AI | USD 74.70 | USD 149.10 | USD 616.50 |
For the support chatbot, the cheapest price (Google Vertex AI, USD 69.84 a month) saves USD 26.46 against the median (USD 96.30). For the coding agent the gap is USD 826.20 a month (63% lower).
Which Kimi K2 Thinking provider is cheapest for your traffic?
Amazon Bedrock, Featherless, Google Vertex AI, Microsoft Azure and Novita AI have the lowest input price (USD 0.60) and GMI Cloud the lowest output price (USD 1.20). They cost the same at 6.5 input tokens per output token. With more input than that, as with long prompts or retrieval, Amazon Bedrock, Featherless, Google Vertex AI, Microsoft Azure and Novita AI are cheaper; with less, as with long answers, GMI Cloud is.
| Provider | No cache hits | 50% cached | 90% cached |
|---|---|---|---|
| GMI Cloud (no cache price) | USD 0.90 cheapest | USD 0.90 | USD 0.90 |
| Amazon Bedrock (no cache price) | USD 1.075 | USD 1.075 | USD 1.075 |
| Featherless (no cache price) | USD 1.075 | USD 1.075 | USD 1.075 |
| Google Vertex AI | USD 1.075 | USD 0.8725 cheapest | USD 0.7105 cheapest |
| Microsoft Azure (no cache price) | USD 1.075 | USD 1.075 | USD 1.075 |
| Novita AI | USD 1.075 | USD 0.9063 | USD 0.7712 |
USD per 1M tokens, 3 input tokens to 1 output token, each provider's cheapest standard row. A provider without a cached price bills cached input at the input price. Any charge for writing to the cache, which some providers add, is not included.
2 of 6 providers publish a cached-input price below the input price: Google Vertex AI (90% off), Novita AI (75% off). GMI Cloud, Amazon Bedrock, Featherless and Microsoft Azure publish none, so cached tokens there bill at the input price.
Context windows differ: 262,144 tokens at GMI Cloud and Novita AI; 32,768 tokens at Featherless. Amazon Bedrock, Google Vertex AI and Microsoft Azure do not state one. The cheapest row is not always the one that fits a long prompt.
What does Kimi K2 Thinking cost in other languages?
The same text needs a different number of tokens in each language. With the Kimi K3 tokenizer, sending 1M characters costs this much at the cheapest standard input price we found for Kimi K2 Thinking:
| Content | Input per 1M characters | Cheapest provider |
|---|---|---|
| English | USD 0.130 | Amazon Bedrock |
| Arabic MSA | USD 0.288 | Amazon Bedrock |
| Hindi | USD 0.341 | Amazon Bedrock |
| Chinese | USD 0.327 | Amazon Bedrock |
| Code | USD 0.156 | Amazon Bedrock |
From BroadRouter's real cost by language dataset, which covers 15 languages and content types for every model we could match to a measured tokenizer.
Estimate your own bill
- LLM API cost calculator with Kimi K2 Thinking preloaded: your traffic and cache rate, against any of 428 models.
- Kimi token counter: count the tokens in your real prompts. Tokenizers differ, so the same text costs a different number of tokens on different models.
- Prompt caching calculator: what a higher cache-hit rate saves.
How does Kimi K2 Thinking compare with other Moonshot AI models?
The Moonshot AI models whose cheapest standard price is closest to Kimi K2 Thinking's (USD 0.90 blended):
| Model | Cheapest input | Cheapest output | Max context | Providers |
|---|---|---|---|---|
| Kimi K2.5 | USD 0.45 | USD 2.25 | 262,144 | 8 |
| Kimi K2 Instruct | USD 0.57 | USD 2.30 | 131,072 | 2 |
| Kimi K2 Instruct 0905 | USD 0.60 | USD 2.50 | 262,144 | 2 |
| Kimi K2.6 | USD 0.39 | USD 1.80 | 262,144 | 13 |
| Kimi K2.7 Code | USD 0.60 | USD 3 | 262,144 | 12 |
| Kimi K3 | USD 1.85 | USD 9.50 | 1,048,576 | 18 |
Priced closest from other makers: GLM 4.5V (Z.ai, USD 0.90) · NVIDIA Nemotron 3 Ultra (NVIDIA, USD 0.9125) · DeepSeek R1 (DeepSeek, USD 0.92).
Compare Kimi K2 Thinking with
- Kimi K2 Thinking vs Kimi K2: API price side by side
Also compared: vs Kimi K2 Instruct.
Every Kimi K2 Thinking price comparison · All model-vs-model price comparisons
All 428 models: API prices compared · Cheapest LLM APIs right now · LLM API price index (open data) · Download all prices (CSV, JSON) · Price change log
Questions
How much does the Kimi K2 Thinking API cost per token?
From USD 0.60 per 1M input tokens (USD 0.0000006 per token) and USD 1.20 per 1M output tokens (USD 0.0000012 per token) at the lowest standard prices we found, which can come from different providers. The median of 6 standard prices is USD 0.60 input and USD 2.50 output. Each price was read from the provider's own pricing page; the source is linked in its row.
Which provider is cheapest for Kimi K2 Thinking?
Among the 6 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest is GMI Cloud, ranked on a blend of 3 input tokens to 1 output token (USD 0.90 per 1M). For input alone the lowest price is USD 0.60 (Amazon Bedrock and 4 others, tied); for output alone, USD 1.20 (GMI Cloud). Batch, flex, priority and fast tiers are shown in the table but not ranked.
Does Moonshot AI sell Kimi K2 Thinking on its own API?
Moonshot AI publishes no price for Kimi K2 Thinking on its own API that we could find, so there is no maker price to compare. The lowest-priced host we checked is GMI Cloud at USD 0.80 / USD 1.20 per 1M tokens.
Does Kimi K2 Thinking have cached input pricing?
2 of 6 standard prices include one. The lowest is USD 0.06 per 1M at Google Vertex AI; the median is USD 0.105. Providers without one bill cached input at the input price. What cached input means.
What is the Kimi K2 Thinking context window?
Up to 262,144 tokens. 1 standard price comes with less, down to 32,768 tokens (Featherless). Check the context column before you pick the cheapest row for long prompts. Context window calculator.
How much does a coding agent cost per month on Kimi K2 Thinking?
For 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning, over 30 days: USD 478.80 a month at Google Vertex AI, USD 1,305 at the median price and USD 1,548 at GMI Cloud. Try your own numbers.
Is Kimi K2 Thinking free to use through an API?
No provider in our data lists a free price for Kimi K2 Thinking as of 2026-11-15. Free tiers and credits by provider (checked 2026-11-14) lists general offers that do not name a model; all free LLM API models are on one page.