Qwen3 30B A3B

Apr 2025
Qwen3 30B A3B
Compare in
on
—per task?
$0.12 / $0.50per 1M in / out?
119 tok/sfirst token in 0.6 s?
33Kcontext

Rankingin non-thinking mode?

Strong at

C
Tool calling#182 of 242Reasoning#288 of 334Science#254 of 312Knowledge#282 of 318Instructions#202 of 245

Alternatives?

ModelScoreCostPer taskWhat changes
CheaperClaude Haiku 5.5max×6.6-7%$0.21Remove sampling parameters+5BetterGPT-6 Lunalow×3.3-7%$0.005Function calling only with none · Chat Completions+3OpenQwen3.8 27Bxhigh×5.1+345%$1.01max_tokens range · QwenCloud+4CheaperMiMo v2.6 FlashthinkingOpen×5.7-19%$0.062Memory (4-bit): 18 GB → 186 GB+4CheaperAgnes 2.5 Pro Beta×5.3-30%$0.10 / $0.30

Run it yourself

estimate
Fits on✗ 8 GB (phone, laptop)✗ 16 GB✓ 24 GB GPU · compressed✓ 48 GB GPU · 8-bit✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use

History

2 versions
7.588.599.510Jul 25Release dateAA Intelligence Index 4.3.2Qwen3 30B A3B2507
7.588.599.510Jul 25Release dateAA Intelligence Index 4.3.2Qwen3 30B A3B2507

Benchmarks

all categories
Chat & Knowledge
Humanity's Last Exam4.6#291
AA Omniscience Index-66#284
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈6.6#284
AA Omniscience · Accuracy12#268
AA Omniscience · Hallucination Rate88.5#234
GPQA Diamond51.5#257
AIME 2025 · AA archived21.7#114
Coding
Terminal-Bench Hard (version unstated)6.8#144
LiveCodeBench · AA archived32.2#114
Agents & Tool Use
τ²-bench · AA22.2#184
Writing & Language
IFBench (instructions)31.9#204

FAQ

How much does Qwen3 30B A3B cost?›

$0.12 per 1M input tokens and $0.50 per 1M output tokens (cheapest host: DeepInfra).

Is Qwen3 30B A3B free to use?›

There is no free API; the cheapest paid price is $0.12 per 1M input tokens. The weights are open, so you can run it yourself.

How good is Qwen3 30B A3B, and what is it best at?›

#288 of 334 in Chat & Knowledge, #142 of 239 in Coding, #182 of 242 in Agents & Tool Use, #141 of 195 in Writing & Language.

Is Qwen3 30B A3B better than Step3 VL 10B?›

On AA Intelligence Index 4.3.2, Qwen3 30B A3B scores 7.6 vs 7.7 — about the same.

How can I use Qwen3 30B A3B?›

Through the API of DeepInfra.

What is the context window of Qwen3 30B A3B?›

33K tokens, with up to 16K output tokens.

When was Qwen3 30B A3B released? Is there a newer version?›

Released 2025-04-28. It has been superseded by Qwen3 30B A3B 2507 (Reasoning).

Is Qwen3 30B A3B open source?›

Yes, the weights are open (Apache-2.0 licence).

Can I run Qwen3 30B A3B locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

Can Qwen3 30B A3B read images?›

No, it accepts text only.

What is Qwen3 30B A3B?›

Qwen3 30B A3B is an open-weights language model from Alibaba (Qwen), released Apr 2025.

What are the best alternatives to Qwen3 30B A3B?›

Claude Haiku 5.5 (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).