modelgrep
Q

Qwen: Qwen3 Coder 30B A3B Instruct

qwen/qwen3-coder-30b-a3b-instruct

119th smartest of 180Cheaper than 90% of paidToolsJSON
Use via OpenRouter ↗
Intelligence
20.0
119th of 180
Design Elo
1132
Website
Speed
74
118th fastest
Latency
875ms
first token
Input price
$0.070
31st cheapest
Context
160K
33K max out

How it compares

Smarter than34%
of all ranked models
Faster than62%
of all ranked models
Cheaper than90%
of all ranked models

Overview

Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use. Built on the...

Benchmarks

independent · via OpenRouter
Artificial Analysis33th percentile
Intelligence Index
20.0
Coding Index
19.4
Agentic Index
21.1
GPQA Diamond
52%
Humanity's Last Exam
4%
SciCode
28%
Tau²-Bench (agentic)
35%
Design Arena · Elo492 tournaments
Website
1132
Data Viz
1128
UI Component
1103

Providers & pricing (4)

ProviderIn $/MOut $/MUptime
Novitafp8$0.070$0.270100%
SiliconFlowfp8$0.070$0.28097.4%
Amazon Bedrock$0.150$0.600
Alibaba$0.292$1.46

Specifications

Context window160K
Max output33K
Knowledge cutoffJun 2025
Input modalitiestext
Output modalitiestext
Prompt caching
Cache read price
ModeratedNo

Qwen3 Coder 30B A3B Instruct FAQ

How much does Qwen3 Coder 30B A3B Instruct cost?

Qwen3 Coder 30B A3B Instruct costs $0.070 per million input tokens and $0.270 per million output tokens via OpenRouter, making it 31st cheapest of 298 paid models.

How smart is Qwen3 Coder 30B A3B Instruct?

Qwen3 Coder 30B A3B Instruct scores 20.0 on the Artificial Analysis Intelligence Index, ranking 119th of 180 benchmarked models, with a GPQA Diamond score of 52%.

How fast is Qwen3 Coder 30B A3B Instruct?

Qwen3 Coder 30B A3B Instruct generates around 74 tokens per second with 875ms time-to-first-token (p50), the 118th fastest tracked model.

What is Qwen3 Coder 30B A3B Instruct's context window?

Qwen3 Coder 30B A3B Instruct supports a 160K-token context window and can output up to 33K tokens. It accepts text input.

Compare head-to-head