Qwen3 Max
AlibabaQwenProprietary
Description
Qwen3 Max is the flagship model in the Qwen3 series, designed for maximum performance across all tasks. It excels in reasoning, mathematics, coding, and complex problem-solving while maintaining strong multilingual capabilities. Optimized for production use cases requiring the highest quality outputs.
Release Date
2025-09-23
Parameters
1.0T
Context Length
262K
Modalities
text
Capability Radar
44
general
45
coding
80
reasoning
50
scienceest.
70
agents
0
multimodal
Science uses a reasoning proxy when dedicated science benchmarks are unavailable.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agents & Tools | 7 | 75.0 | LS |
| Code Ranking | 126 | 54.0 | AA |
| General Ranking | 129 | 62.0 | AA |
| Math Reasoning | 77 | 82.0 | AA |
| Science | 138 | 56.0 | AA |
Benchmark Scores (LLM Stats)
Agents
t2-bench
74.8%SR
Biology
GPQA
62.0%SR
Chemistry
SuperGPQA
65.1%SR
Code
SWE-Bench Verified
69.6%SR
General
LiveCodeBench v6
69.0%SR
Math
AIME 2025
81.6%SR
AA Evaluation Indices
Math Index80.7
Intelligence Index31.4
Coding Index26.4
Mmlu Pro0.8
Aime 250.8
Livecodebench0.8
Gpqa0.8
Tau20.7
Lcr0.5
Ifbench0.4
Scicode0.4
Terminalbench Hard0.2
Hle0.1
LLM Stats Category Scores
Tool Calling70
Agents70
Code70
Economics70
Finance70
Frontend Development70
General70
Healthcare70
Legal70
Math70
Reasoning70
Biology60
Chemistry60
Physics60
Pricing
Input Price$1.655 / 1M tokens
Output Price$7.225 / 1M tokens
Blended Price (3:1)$3.047 / 1M tokens
Speed
Tokens/sec33.0 tokens/s
Time to First Token1.93s
Time to Answer1.93s
Available Providers
(LS internal units)| Provider | Input Price | Output Price |
|---|---|---|
| Novita | 500K | 5.0M |