BroadRouter

Kimi K2.5 API pricing

Updated

Among the 8 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest blended standard price for Kimi K2.5 is SiliconFlow at USD 0.45 per 1M input tokens and USD 2.25 per 1M output tokens. The median across 8 standard prices is USD 0.60 / USD 3. Each price is read from the provider's own pricing page.

Moonshot AI · slug kimi-k2.5 · up to 262,144 tokens of context · 8 providers with a standard price. Every price is read from the provider's own pricing page and linked in its row, checked 2026-11-15. No aggregator feed. Download all prices (every provider and model, with sources, CSV or JSON). How it compares with the market: the LLM API price index. Archived list prices since 2026-02-01 (8 archived price points, 0 list-price changes): LLM API price history.

Does Moonshot AI sell Kimi K2.5 directly?

Moonshot AI publishes no price for Kimi K2.5 on its own API that we could find, so this page compares hosts only.

Kimi K2.5 model facts

Facts from Moonshot AI's model card for moonshotai/Kimi-K2.5 and the Hugging Face model API, read 2026-11-16. BroadNet did not verify them, and hosts may serve the model at a different precision. BroadRouter is BroadNet's site.

Kimi K2.5: facts as stated by the model card and the Hugging Face model API, with sources
FactAs statedSource
Open weightsYes. The weights are in a public Hugging Face repository (moonshotai/Kimi-K2.5) with no access gate.Hugging Face model API, checked 2026-11-16
Quote
“"private": false, "gated": false, 64 .safetensors files listed”
LicenceCustom licence: modified-mitMoonshot AI model card metadata, checked 2026-11-16
Quote
“license: other; license_name: modified-mit”
Total parameters1TMoonshot AI model card, checked 2026-11-16
Quote
“Total Parameters | 1T”
Active parameters per token32BMoonshot AI model card, checked 2026-11-16
Quote
“Activated Parameters | 32B”
Context length256KMoonshot AI model card, checked 2026-11-16
Quote
“Context Length | 256K”
Input and outputImage and text in, text out (the card's task tag is image-text-to-text)Moonshot AI model card metadata, checked 2026-11-16
Quote
“pipeline_tag: image-text-to-text”
Tool callingYesMoonshot AI model card, checked 2026-11-16
Quote
“K2.5 shares the same design of Interleaved Thinking and Multi-Step Tool Call as K2 Thinking.”
First published on Hugging Face (not necessarily the release date)2026-01-01Hugging Face model API, checked 2026-11-16
Quote
“"createdAt": "2026-01-01T06:06:03.000Z"”

Kimi K2.5 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 61% above the cheapest.

Kimi K2.5 API prices by provider
#ProviderInputCached inputOutputBlended 3:1ContextPrecisionSource
1SiliconFlow cheapest
Notes and source quote

Context shown as 262K.

“Kimi-K2.5 262K $0.45 $0.07 $2.25”
USD 0.45USD 0.07USD 2.25USD 0.90262,144not statedSiliconFlow page
checked 2026-11-14
re-checked 2026-11-16
2Amazon Bedrock
Notes and source quote

Table headed "Regions: US East (N. Virginia), US East (Ohio) and US West (Oregon)".

“Kimi K2.5 $ 0.60 $ 3.00”
USD 0.60not publishedUSD 3USD 1.20not statednot statedAmazon Bedrock page
checked 2026-11-14
re-checked 2026-11-16
3Cloudflare Workers AI
@cf/moonshotai/kimi-k2.5
Notes and source quote

Cloudflare publishes this USD per-token price alongside the neuron price.

“@cf/moonshotai/kimi-k2.5 $0.600 per M input tokens $0.100 per M cached input tokens $3.000 per M output tokens”
USD 0.60USD 0.10USD 3USD 1.20not statednot statedCloudflare Workers AI page
checked 2026-11-14
re-checked 2026-11-16
4GMI Cloud
moonshotai/Kimi-K2.5
Notes and source quote

No cache-read price published.

“Price (per 1M tokens) Input $0.60 Output $3.00”
USD 0.60not publishedUSD 3USD 1.20262,144FP8GMI Cloud page
checked 2026-11-15
re-checked 2026-11-16
5Microsoft Azure
Notes and source quote

Foundry Models serverless (pay-as-you-go), Global deployment.

“Kimi K2.5 Thinking Global $0.60 $0.10 $3”
USD 0.60USD 0.10USD 3USD 1.20not statednot statedMicrosoft Azure page
checked 2026-11-14
re-checked 2026-11-16
6Novita AI
moonshotai/kimi-k2.5
Notes and source quote
“Kimi K2.5 256K $0.6 /Mt · Cache Read $0.1 /Mt $3 /Mt”
USD 0.60USD 0.10USD 3USD 1.20262,144not statedNovita AI page
checked 2026-11-14
re-checked 2026-11-16
7Venice
kimi-k2-5
Notes and source quote

Marked "Retiring" by Venice.

“| Kimi K2.5 (Retiring) | `kimi-k2-5` | $0.56 | $3.50 | $0.22 | — | Private |”
USD 0.56USD 0.22USD 3.50USD 1.295256,000not statedVenice page
checked 2026-11-15
re-checked 2026-11-16
8Featherless
moonshotai/Kimi-K2.5
Notes and source quote

Developer plan (per-token billing, from 50 USD a month in credits).

“PRICING: INPUT $0.8 / CACHED $0.154 / OUTPUT $3.4”
USD 0.80USD 0.154USD 3.40USD 1.45262,144FP8Featherless page
checked 2026-11-15
re-checked 2026-11-16

Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).

What does Kimi K2.5 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.
Monthly cost of three workloads on Kimi K2.5, by provider
ProviderSupport chatbotRAG over documentsCoding agent (one developer)
SiliconFlowUSD 60.48USD 114.39USD 431.10
Amazon BedrockUSD 108.00USD 189.00USD 1,350
Cloudflare Workers AIUSD 81.00USD 153.00USD 585.00
GMI CloudUSD 108.00USD 189.00USD 1,350
Microsoft AzureUSD 81.00USD 153.00USD 585.00
Novita AIUSD 81.00USD 153.00USD 585.00
VeniceUSD 95.04USD 162.42USD 802.80
FeatherlessUSD 98.32USD 196.49USD 757.62

For the support chatbot, the cheapest price (SiliconFlow, USD 60.48 a month) saves USD 27.54 against the median (USD 88.02). For the coding agent the gap is USD 240.21 a month (36% lower).

Which Kimi K2.5 provider is cheapest for your traffic?

SiliconFlow has the lowest input price (USD 0.45) and the lowest output price (USD 2.25), so it is the cheapest standard price at every mix of input and output tokens, before cache discounts.

Kimi K2.5: price per 1M tokens at 3 input to 1 output, by share of input served from cache
ProviderNo cache hits50% cached90% cached
SiliconFlowUSD 0.90 cheapestUSD 0.7575 cheapestUSD 0.6435 cheapest
Amazon Bedrock (no cache price)USD 1.20USD 1.20USD 1.20
Cloudflare Workers AIUSD 1.20USD 1.012USD 0.8625
GMI Cloud (no cache price)USD 1.20USD 1.20USD 1.20
Microsoft AzureUSD 1.20USD 1.012USD 0.8625
Novita AIUSD 1.20USD 1.012USD 0.8625

USD per 1M tokens, 3 input tokens to 1 output token, each provider's cheapest standard row; 2 dearer providers are in the price table above. 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.

6 of 8 providers publish a cached-input price below the input price: SiliconFlow (84% off), Cloudflare Workers AI (83% off), Microsoft Azure (83% off), Novita AI (83% off), Venice (61% off), Featherless (81% off). Amazon Bedrock and GMI Cloud publish none, so cached tokens there bill at the input price.

Context windows differ: 262,144 tokens at SiliconFlow, GMI Cloud, Novita AI and Featherless; 256,000 tokens at Venice. Amazon Bedrock, Cloudflare Workers AI and Microsoft Azure do not state one. The cheapest row is not always the one that fits a long prompt.

What does Kimi K2.5 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 K2.5:

ContentInput per 1M charactersCheapest provider
EnglishUSD 0.098SiliconFlow
Arabic MSAUSD 0.216SiliconFlow
HindiUSD 0.255SiliconFlow
ChineseUSD 0.245SiliconFlow
CodeUSD 0.117SiliconFlow

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.

Published benchmark results

BroadNet did not run these benchmarks; results are as published by the Moonshot AI model card and may use different settings. BroadRouter is BroadNet's site.

Published benchmark results for Kimi K2.5, with sources
BenchmarkResultSetting or noteSource
GPQA-Diamond87.6Thinking modeMoonshot AI model card, checked 2026-11-16
MMLU-Pro87.1Thinking modeMoonshot AI model card, checked 2026-11-16
SWE-Bench Verified76.8Thinking modeMoonshot AI model card, checked 2026-11-16
LiveCodeBench (v6)85.0Thinking modeMoonshot AI model card, checked 2026-11-16

Estimate your own bill

The Moonshot AI models whose cheapest standard price is closest to Kimi K2.5's (USD 0.90 blended):

Moonshot AI models priced closest to Kimi K2.5
ModelCheapest inputCheapest outputMax contextProviders
Kimi K2 ThinkingUSD 0.80USD 1.20262,1446
Kimi K2 InstructUSD 0.57USD 2.30131,0722
Kimi K2 Instruct 0905USD 0.60USD 2.50262,1442
Kimi K2.6USD 0.39USD 1.80262,14413
Kimi K2.7 CodeUSD 0.60USD 3262,14412
Kimi K3USD 1.85USD 9.501,048,57618

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 K2.5 with

Also compared: vs GLM 5.3 · vs Kimi K2 Instruct · vs Kimi K2.6.

Every Kimi K2.5 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 K2.5 API cost per token?

From USD 0.45 per 1M input tokens (USD 0.00000045 per token) and USD 2.25 per 1M output tokens (USD 0.00000225 per token) at the lowest standard prices we found. The median of 8 standard prices is USD 0.60 input and USD 3 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 K2.5?

Among the 8 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest is SiliconFlow, ranked on a blend of 3 input tokens to 1 output token (USD 0.90 per 1M). It also has the lowest input and the lowest output price. Batch, flex, priority and fast tiers are shown in the table but not ranked.

Does Moonshot AI sell Kimi K2.5 on its own API?

Moonshot AI publishes no price for Kimi K2.5 on its own API that we could find, so there is no maker price to compare. The lowest-priced host we checked is SiliconFlow at USD 0.45 / USD 2.25 per 1M tokens.

Does Kimi K2.5 have cached input pricing?

6 of 8 standard prices include one. The lowest is USD 0.07 per 1M at SiliconFlow; the median is USD 0.10. Providers without one bill cached input at the input price. What cached input means.

What is the Kimi K2.5 context window?

Up to 262,144 tokens. 1 standard price comes with less, down to 256,000 tokens (Venice). 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 K2.5?

For 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning, over 30 days: USD 431.10 a month at SiliconFlow, USD 671.31 at the median price and USD 1,350 at GMI Cloud. Try your own numbers.

Is Kimi K2.5 free to use through an API?

No provider in our data lists a free price for Kimi K2.5 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.

Is Kimi K2.5 open source, with open weights, and what is its licence?

Moonshot AI publishes the weights on Hugging Face as moonshotai/Kimi-K2.5, in a public repository with no access gate. The licence, as the Moonshot AI model card metadata states it: Custom licence: modified-mit. Open weights means you can download and run the model yourself; what you may do with it, including commercial use, is set by that licence, so read it in full (checked 2026-11-16). These facts come from Moonshot AI's card and the Hugging Face model API; BroadNet did not verify them, and hosts may serve the model at a different precision.

How many parameters does Kimi K2.5 have?

Total parameters, as the Moonshot AI model card states them: 1T. Active parameters per token: 32B. (checked 2026-11-16; BroadNet did not verify these figures.)

Does Kimi K2.5 support tool calling?

Yes, according to the Moonshot AI model card: “K2.5 shares the same design of Interleaved Thinking and Multi-Step Tool Call as K2 Thinking.” (checked 2026-11-16). Whether the tools parameter works through an API also depends on the provider serving it, so check that provider's docs. BroadNet did not test it.

How does Kimi K2.5 score on benchmarks?

As published, the Moonshot AI model card reports 87.6 on GPQA-Diamond, 87.1 on MMLU-Pro and 76.8 on SWE-Bench Verified (checked 2026-11-16). BroadNet did not run these benchmarks, and sources may use different settings, so the figures are not directly comparable. Every published result we found, with its setting and source.

Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity