Qwen3, the latest generation in the Qwen large language model series, features both dense and mixture-of-experts (MoE) architectures to excel in reasoning, multilingual support, and advanced agent tasks. Its unique...
Tested on 9 benchmarks · BenchGecko score 34.0. Top scores: Chatbot Arena Elo — Overall (1326.5%), Lech Mazur Writing (75.3%), OTIS Mock AIME 2024-2025 (62.7%).
Qwen2.5 7B Instruct scores 34.8 (102% as good) at $0.10/1M input · 17% cheaper
Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.
Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.
Writing quality evaluation by Lech Mazur. Tests prose quality, coherence, and stylistic ability.
Graduate-level science questions written by PhD experts. Diamond subset contains questions where experts disagree, testing deep understanding.
LiveBench fiction analysis. Tests literary comprehension and creative text understanding.
- Typetext
- Context131K tokens (~66 books)
- ReleasedApr 2025
- LicenseOpen Source
- StatusActive
- Cost / Message~$0.001
Frequently Asked Questions
Related Models
QwQ 32B · Alibaba QwenGPT-4o-mini (2024-07-18) · OpenAILLaMA-13B · MetaGPT-4.5 · OpenAIGemma 2B · Google DeepMindKey facts · as of 2026-10-05
- Qwen3 30B A3B by Alibaba Qwen. BenchGecko score 34.0, rank 229 of 312 scored models (normalized average of public benchmark scores).
- List price $0.12 input · $0.50 output per 1M tokens (as of 2026-10-05).
- Sold by 2 providers (as of 2026-10-05): DeepInfra (fp8) $0.12 in / $0.50 out · Alibaba $0.13 in / $0.52 out. Every provider
How to cite · data as of 2026-10-05
Qwen3 30B A3B · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/qwen3-30b-a3b
Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP