The AI arena is free today

Open Superagent

Qwen3 14B vs Qwen3 30B A3B

Qwen3 14B and Qwen3 30B A3B are closely matched at 18.2 and 17.4 on the LLM Stats Score. Qwen3 14B is 1.0x cheaper per token.

Alibaba Cloud / Qwen Team · Alibaba Cloud / Qwen Team · Updated for 2026

Which is better?

Qwen3 14B and Qwen3 30B A3B are closely matched on the overall LLM Stats Score at 18.2 and 17.4.

The models split the 4 individual benchmarks reported for both models evenly.

Qwen3 30B A3B also accepts a larger context window (128,000 input tokens), making it the stronger choice for long documents and large codebases.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose Qwen3 14B

  • you want predictable pricing at $0.12/M input and $0.24/M output

Choose Qwen3 30B A3B

  • you process long inputs — it offers a 128,000 token context window
  • you want the most recent training data — it shipped Apr 2025

At a glance

The differences that matter most.

Core performance indexes
18.2
#212
17.4
#215
19.6
#196
17.7
#206
Cost, coverage & limits
Benchmark wins
2 of 4
2 of 4
Input price
$0.12 / M
$0.10 / M
Output price
$0.24 / M
$0.30 / M
Context window
40,960
128,000

Capability indexes

Additional strengths measured across groups of related public benchmarks

1 shared
Index
Qwen3 14B
Qwen3 30B A3B
21.4#145
19.6#163
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

17 reported for Qwen3 14B · 8 for Qwen3 30B A3B

4 shared

Qwen3 14B outperforms in 2 benchmarks (Arena Hard, Multi-IF), while Qwen3 30B A3B is better at 2 benchmarks (AIME 2024, AIME 2025).

Both models are evenly matched across the benchmarks.

Tue Sep 08 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Pricing Analysis

Price comparison per million tokens

Qwen3 14B costs less

For input processing, Qwen3 14B ($0.12/1M tokens) is 1.2x more expensive than Qwen3 30B A3B ($0.10/1M tokens).

For output processing, Qwen3 14B ($0.24/1M tokens) is 1.3x cheaper than Qwen3 30B A3B ($0.30/1M tokens).

In conclusion, Qwen3 30B A3B is more expensive than Qwen3 14B.*

* Using a 3:1 ratio of input to output tokens

Lowest available price from all providers
Tue Sep 08 2026 • llm-stats.com
Alibaba Cloud / Qwen Team
Qwen3 14B
Input tokens$0.12
Output tokens$0.24
Best providerDeepinfra
Alibaba Cloud / Qwen Team
Qwen3 30B A3B
Input tokens$0.10
Output tokens$0.30
Best providerDeepinfra
Notice missing or incorrect data?Start an Issue

Model Size

Parameter count comparison

16.5B diff

Qwen3 30B A3B has 16.5B more parameters than Qwen3 14B, making it 117.9% larger.

Alibaba Cloud / Qwen Team
Qwen3 14B
14.0Bparameters
Alibaba Cloud / Qwen Team
Qwen3 30B A3B
30.5Bparameters
14.0B
Qwen3 14B
30.5B
Qwen3 30B A3B

Context Window

Maximum input and output token capacity

Qwen3 30B A3B accepts 128,000 input tokens compared to Qwen3 14B's 40,960 tokens. Qwen3 30B A3B can generate longer responses up to 128,000 tokens, while Qwen3 14B is limited to 40,960 tokens.

Alibaba Cloud / Qwen Team
Qwen3 14B
Input40,960 tokens
Output40,960 tokens
Alibaba Cloud / Qwen Team
Qwen3 30B A3B
Input128,000 tokens
Output128,000 tokens
Tue Sep 08 2026 • llm-stats.com

License

Usage and distribution terms

Both models are licensed under Apache 2.0.

Both models share the same licensing terms, providing consistent usage rights.

Qwen3 14B

Apache 2.0

Open weights

Qwen3 30B A3B

Apache 2.0

Open weights

Release Timeline

When each model was launched

Qwen3 14B was released on 2025-04-28, while Qwen3 30B A3B was released on 2025-04-29.

Qwen3 30B A3B is 0 month newer than Qwen3 14B.

Qwen3 14B

Apr 28, 2025

1.4 years ago

Qwen3 30B A3B

Apr 29, 2025

1.4 years ago

1d newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Provider Availability

Qwen3 14B is available from DeepInfra. Qwen3 30B A3B is available from DeepInfra, Novita, Fireworks.

Qwen3 14B

deepinfra logo
Deepinfra
Input Price:Input: $0.12/1MOutput Price:Output: $0.24/1M

Qwen3 30B A3B

deepinfra logo
Deepinfra
Input Price:Input: $0.10/1MOutput Price:Output: $0.30/1M
novita logo
Novita
Input Price:Input: $0.10/1MOutput Price:Output: $0.44/1M
fireworks logo
Fireworks
Input Price:Input: $0.89/1MOutput Price:Output: $0.89/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against Qwen3 14B and Qwen3 30B A3B side-by-side, then vote on the output you prefer.

Qwen3 14B
✓ Preferred
Qwen3 30B A3B
Open in Playground

FAQ

Common questions about Qwen3 14B vs Qwen3 30B A3B.

Which is better, Qwen3 14B or Qwen3 30B A3B?

Qwen3 14B and Qwen3 30B A3B are closely matched on the LLM Stats Score at 18.2 and 17.4. Qwen3 14B is made by Alibaba Cloud / Qwen Team and Qwen3 30B A3B is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Qwen3 14B compare to Qwen3 30B A3B in benchmarks?

Qwen3 14B scores MATH-500: 96.8%, Arena Hard: 91.7%, RULER: 90.1%, AutoLogi: 89.2%, MMLU-Redux: 88.6%. Qwen3 30B A3B scores Arena Hard: 91.0%, AIME 2024: 80.4%, LiveBench: 74.3%, Multi-IF: 72.2%, AIME 2025: 70.9%.

Is Qwen3 14B cheaper than Qwen3 30B A3B?

Qwen3 30B A3B is 1.2x cheaper for input tokens. Qwen3 14B costs $0.12/M input and $0.24/M output via deepinfra. Qwen3 30B A3B costs $0.10/M input and $0.30/M output via deepinfra.

What are the context window sizes for Qwen3 14B and Qwen3 30B A3B?

Qwen3 14B supports 41K tokens and Qwen3 30B A3B supports 128K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Qwen3 14B and Qwen3 30B A3B?

Key differences include LLM Stats Score (18.2 vs 17.4), context window (41K vs 128K), input pricing ($0.12 vs $0.10/M). See the full comparison above for benchmark-by-benchmark results.