Last verified
KWAIPILOT KAT-CODER256K CONTEXTFAST TIERCNY SOURCE

KAT-Coder-Air V2.5 API Pricing

KAT-Coder-Air V2.5 is the fast tier of Kwaipilot's coding line, served on Kuaishou's StreamLake API. At CNY 1/M input and CNY 4/M output ($0.14 and $0.56 at the project rate of 7.10) it lands at an effective $0.09/M - a fifth of the Pro tier, with the same 256K context and 80K output ceiling.

Input - per 1M tokens
$0.14/M
CNY 1 per 1M at 7.10 list price
Output - per 1M tokens
$0.56/M
CNY 4 per 1M at 7.10 list price
Cached input - per 1M tokens
$0.03/M
Cache writes free -80%
Effective - agentic blend
$0.09/M
92/8 split - 82% cache
§ 01 / TERMINAL

Run the numbers.

Live calculator pre-loaded with current KAT-Coder-Air V2.5 rates. Tweak the workload split, then share the URL to share the calculation.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
Words equivalent (English)
Effective rate
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

Real-world presets.

§ 03 / TOKENIZER

Paste text. See tokens. See cost.

Estimate · kwaipilot-tokenizer-estimate · ≈3.85 chars/token Auto-counts as you type

This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.

Characters 615
Words 103
Tokens (estimated) 160 tokens
Cost as input · uncached $0.00002 USD
Cost as output · uncached $0.00009 USD
Cost as cached input $0.000005 USD
§ 04 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
KAT-Coder-Air V2.5 Current $0.14 cache $0.03 $0.56 $0.09 agentic 92/8 256K High-volume agent and Claw loops
ERNIE 4.5 Turbo $0.12 cache $0.03 $0.47 $0.08 cheapest verified peer 128K Budget Chinese-language work
Hunyuan HY3 $0.14 cache $0.04 $0.56 $0.09 near-identical rate card 256K Tencent Cloud agent workloads
KAT-Coder-Pro V2 $0.30 cache $0.06 $1.18 $0.19 previous Pro, still listed not listed Stronger tier at the old Pro rate
LongCat-2.0 $0.70 cache $0.01 $2.82 $0.35 1M-context agentic peer 1M Agentic coding with 1M context
KAT-Coder-Pro V2.5 $0.70 cache $0.14 $2.82 $0.45 same family, flagship 256K Repo-scale agentic coding tasks

Frequently asked.

Practical KAT-Coder-Air V2.5 pricing questions, with the CNY list price separated from workload assumptions.

Q · 01 What does KAT-Coder-Air V2.5 cost? +
The StreamLake price table lists CNY 1 per 1M input tokens, CNY 4 per 1M output tokens and CNY 0.2 per 1M cached input tokens, with cache writes free. At the project rate of 7.10 CNY/USD that is $0.140845, $0.56338 and $0.028169.
Q · 02 How much cheaper is it than the Pro tier? +
Five times. Air V2.5 blends to $0.09/M against $0.45/M for Pro V2.5, on identical 256K context and 80K output limits. The trade is benchmark score: the vendor reports SWE-bench-Pro 42.4% for Air versus 65.2% for Pro.
Q · 03 What is it built for? +
Kwaipilot positions it for Agent and Claw scenarios - high-frequency loops where latency and cost per call dominate. The product page marks its relative latency as fast and lists MCP, function calling, streaming and context caching, the same feature set as the Pro tier.
Q · 04 Are cache writes free here too? +
Yes. The product page marks 缓存写入 (cache write) as 免费 (free) for both V2.5 models. Only reads are billed, at CNY 0.2/M - an 80% discount against a miss. A miss bills at the standard input rate.
Q · 05 Is anything cheaper at this capability? +
Among freshly verified peers, ERNIE 4.5 Turbo is 11% cheaper on the same blend but caps at 128K context. Hunyuan HY3 is a near-perfect price twin - $0.141/$0.563 against this model's $0.141/$0.563 - at the same 256K context, so the choice between them comes down to capability and platform, not rate.
Q · 06 Are the older KAT-Coder models still free? +
Two of them are. The price table still lists KAT-Coder-Air-V1 and KAT-Coder-Exp-72B-1010 as 免费 (free) across input, output and cache, dated October 2025. No published quota was located, so assume rate limits apply.
Q · 07 Where does the API live? +
On StreamLake (万擎 / Vanchin), Kuaishou's own cloud. That console also resells third-party models such as Qwen, GLM and DeepSeek at StreamLake's own rates - but for the KAT-Coder line it is the first-party channel, so these are Kwaipilot's rates.
Q · 08 When was this checked? +
The StreamLake price page and the KAT-Coder product page were both opened on July 22, 2026, and the two agreed on every figure. The price rows themselves are dated 2026/07/07 by the vendor.