Baidu - Closed weights

ERNIE 4.5 Turbo (32K) pricing and specs

Name: ERNIE 4.5 Turbo (32K)
Author: Baidu

ERNIE 4.5 Turbo (32K) is a closed-weights model from Baidu that costs $0.11 per 1M input tokens and $0.45 per 1M output tokens on the first-party API, with a 32K-token context window.

Last verified October 17, 2026 from intl.cloud.baidu.com. All sources

Input / 1M: $0.11
Output / 1M: $0.45
Context: 32K tokens
Weights: Closed

Pricing per 1M tokens

Item	Price / 1M tokens	Applies to
Input	$0.11	Standard requests
Output	$0.45	Standard requests
Cache read	$0.03	Standard requests

Note: Previous-generation model still sold on Baidu AI Cloud international (Qianfan), pricing page updated 2026-09-07. cache read is the listed 'Cache Hit' price; no cache write or batch price listed. Same price as the 128K endpoint, with a shorter window. Context length stated as '32k' (max input 27k), recorded as 32000; max output [2, 12288]. Release date from Baidu's Create 2025 press release (April 25, 2025) announcing ERNIE 4.5 Turbo. Not listed in the Vercel AI Gateway model list on 2026-10-17.

Specs

Context window: 32,000 tokens
Max output: 12,288 tokens
Release date: April 25, 2025
Provider: Baidu
API model id: ernie-4.5-turbo-32k
Weights: Closed weights

What ERNIE 4.5 Turbo (32K) costs in practice

Worked from the standard list rates above, with the math shown so you can swap in your own token counts.

One agent session

1M input tokens and 100K output tokens, no caching.

Input1M x $0.11 = $0.11
Output0.1M x $0.45 = $0.045

$0.155per session

A chat app

10,000 requests a day at 2K input and 500 output tokens each.

Input20M x $0.11 = $2.20
Output5M x $0.45 = $2.25

$4.45per day

$133.50 per 30-day month

The same session at a 90% cache hit

90% of the input is read from the prompt cache. The one-time cache write premium is not included.

Cached input0.9M x $0.03 = $0.027
Uncached input0.1M x $0.11 = $0.011
Output0.1M x $0.45 = $0.045

$0.083per session

Compare with

Model	Input / 1M	Output / 1M	Context	Why listed
ERNIE 4.5 Turbo (32K) (this page)	$0.11	$0.45	32K	-
ERNIE 4.5 Turbo (128K)Baidu	$0.11	$0.45	128K	Similar output price, same provider
Qwen3.8-FlashAlibaba Cloud (Qwen)	$0.15	$0.47	1M	Similar output price
Qwen3.8-Omni-FlashAlibaba Cloud (Qwen)	$0.15	$0.47	1M	Similar output price
Qwen-MT-FlashAlibaba Cloud (Qwen)	$0.16	$0.49	16K	Similar output price
ERNIE 4.5 Turbo VLBaidu	$0.42	$1.27	128K	Same provider
ERNIE 4.5 Turbo VL (32K)Baidu	$0.42	$1.27	32K	Same provider
ERNIE 5.1Baidu	$0.56	$2.53	128K	Same provider
ERNIE 5.0Baidu	$0.84	$3.38	128K	Same provider

Unlimited OCR: Baidu's Open-Source Solution for Long Document ParsingUnlimited OCR is Baidu's open-source vision-language model that parses 100+ page documents in one pass without memory blowup. How it works and when to use it.

FAQ

How much does ERNIE 4.5 Turbo (32K) cost per million tokens?

ERNIE 4.5 Turbo (32K) costs $0.11 per 1M input tokens and $0.45 per 1M output tokens on Baidu's first-party API, per intl.cloud.baidu.com on October 17, 2026. Per 1M tokens, cache reads cost $0.03.

What is ERNIE 4.5 Turbo (32K)'s context window?

ERNIE 4.5 Turbo (32K) has a 32,000-token context window and returns up to 12,288 output tokens per response.

Are ERNIE 4.5 Turbo (32K)'s weights open?

No. ERNIE 4.5 Turbo (32K) is a closed-weights model. You call it through Baidu's API as ernie-4.5-turbo-32k.

When was ERNIE 4.5 Turbo (32K) released?

Baidu released ERNIE 4.5 Turbo (32K) on April 25, 2025.

Sources

Last verified October 17, 2026 from intl.cloud.baidu.com.

Newsletter

One email per week. New model releases, price cuts and what they mean for your bill.