Baidu: ERNIE 4.5 21B A3B Thinking

Public pricingIntelligence 79/100Medium memoryTiefes DenkenTool-Nutzung

Baidu: ERNIE 4.5 21B A3B Thinking ist ein Text-Modell für Programmierung, Debugging und technische Arbeit. Es verbindet starke Coding-Leistung und tiefes Schlussfolgern und Planen, einen Kontext von 131K tokens und ein kostengünstig-Profil für zuverlässige Arbeit über coding, debugging, and technical writing.

Input

$0.07/1M

Output

$0.28/1M

Cached

$0.01/1M

Batch

$0.04/1M

Calculate your ERNIE 4.5 21B A3B Thinking bill.

Set your workload — see cost at your exact volume.

What would ERNIE 4.5 21B A3B Thinking cost you?

Adjust the workload to see your monthly bill.

1,00010,00050,000250,0001M10M

Technical specifications

ERNIE 4.5 21B A3B Thinking at a glance.

Memory

131,072

tokens

Max reply

65,536

tokens

Memory tier

Medium

a long report or a codebase file

Tokenizer

—

Released

Sep 2025

Training cutoff

Feb 2025

Availability

Public pricing

Status

active

What it can do

Capabilities & limits.

  • Understands images
  • Deep step-by-step thinking
  • Uses tools / calls functions
  • Strict JSON output
  • Streams replies
  • Fine-tunable on your data

When to pick ERNIE 4.5 21B A3B Thinking

  • Multi-step reasoning, research agents, or hard math.
  • Agentic workflows that call tools or APIs.
  • High-volume workloads where unit cost matters.

When to look elsewhere

  • Your workload involves images — pick a vision-capable model instead.

Alternatives

Similar models to compare.

FAQ

ERNIE 4.5 21B A3B Thinking — the questions we see most.

Pricing, capabilities, alternatives — generated from the same data that powers the calculator above.

Get instant answers from our AI agent

At a typical workload of 50,000 conversations a month with 1,500-token prompts and 800-token replies, ERNIE 4.5 21B A3B Thinking costs roughly $16 per month. Input is $0.07 /1M tokens and output is $0.28 /1M tokens.
ERNIE 4.5 21B A3B Thinking has a 131,072-token context window (medium memory — a long report or a codebase file). That means you can fit about 24,576 words of input and history in a single call.
Beyond text generation, ERNIE 4.5 21B A3B Thinking supports deep step-by-step reasoning, calling functions / tools, strict JSON output, fine-tuning on your own data. It streams replies by default.
ERNIE 4.5 21B A3B Thinking was released in September 2025, with training data cut off around February 2025.
Models in a similar class include Seed 1.6 Flash, Gemini 2.0 Flash Lite, gpt-oss-safeguard-20b. The "Similar models" section below this FAQ links into each.

Still unsure?

Compare ERNIE 4.5 21B A3B Thinking against 100+ other models.

Open the full wizard — pick a use case, set your usage, and see side-by-side monthly costs in under a minute.

Open the comparison wizard