Alibaba Cloud / Qwen Team model product
3 billion activated parameters.
Updated Aug 12, 2026. Default version: Qwen3 30B A3B
Technical details for the model's default version.
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Benchmark scores for Qwen3 30B A3B.
| Arena Hard | 91.0% | 03 | 26 | 92.0% | C | |
| BFCL | 69.1% | 07 | 11 | 40.0% | C | |
| Multi-IF | 72.2% | 13 | 20 | 36.8% | C | |
| LiveBench | 74.3% | 18 | 38 | 54.0% | C | |
| AIME 2024 | 80.4% | 27 | 53 | 50.0% | C | |
| LiveCodeBench | 62.6% | 31 | 73 | 58.3% | C | |
| AIME 2025 | 70.9% | 87 | 114 | 23.9% | C | |
| GPQA | 65.8% | 147 | 234 | 37.3% | C |
Preference and agent-evaluation results for the default version.
| text | overall | 144 | 1384.2 | 23,836 | N/A | |
| text style control | overall | 160 | 1383.3 | 23,836 | N/A | |
| text | overall | 210 | 1316.9 | 26,561 | N/A | |
| text style control | overall | 222 | 1327.1 | 26,561 | N/A |
Provider-specific output speed and catalog latency for Qwen3 30B A3B. Runtime does not affect the capability score.
| Fireworks | 122.4 tok/s | 0.66 s | 128K | 128K | |
| Novita | 88.84 tok/s | 0.73 s | 128K | 128K | |
| DeepInfra | 82.57 tok/s | 0.84 s | 128K | 128K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Official vendor API pricing appears first, followed by individual provider offers.
| Helicone | qwen3-30b-a3b | global | $0.08 | $0.29 | 41K | |
| NovitaAI | qwen/qwen3-30b-a3b-fp8 | global | $0.09 | $0.45 | 41K | |
| Jiekou.AI | qwen/qwen3-30b-a3b-fp8 | global | $0.09 | $0.45 | 41K | |
| NanoGPT | qwen/qwen3-30b-a3b | global | $0.10 | $0.30 | 41K | |
| 302.AI | qwen3-30b-a3b | global | $0.11 | $1.08 | 128K | |
| Vercel AI Gateway | alibaba/qwen-3-30b | global | $0.12 | $0.50 | 41K | |
| OpenRouter | qwen/qwen3-30b-a3b | global | $0.12 | $0.50 | 131.1K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Available versions of this model. The score column identifies the version used in the overall ranking.
| Qwen3 30B A3B | 32.1 | 30.5B | 128K | 128K | No | Apache 2.0 |
Key information about Qwen3 30B A3B and its available data.
3 billion activated parameters. Features hybrid thinking/non-thinking modes, support for 119 languages, and enhanced agent capabilities. It aims to outperform previous models like QwQ-32B while using significantly fewer activated parameters.
Data as of 2026-08-11.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
Common questions about Qwen3 30B A3B.
Qwen3 30B A3B's default version was released on Apr 29, 2025.
No official standard PAYG price is currently available for Qwen3 30B A3B. The lowest tracked third-party offer starts at $0.08 input and $0.29 output via Helicone.
Qwen3 30B A3B was created by Alibaba Cloud / Qwen Team.
The default version has a 128K token context window.
No. The default version is not marked as having publicly available weights.
8 provider offerings are linked to the default version.
Nearby ranked alternatives include Qwen3.5 4B, Nemotron Nano 9B, Claude Sonnet 3.5.