Qwen3-Max Reasoning: pricing, context window and benchmarks
Qwen3-Max Reasoning is a frontier model from Alibaba Cloud. It offers a 2,000,000-token context window, costs $18.12 per million input tokens and $81.94 per million output tokens, and averages 85.9 across the index's benchmark suite.
Pricing
Qwen3-Max Reasoning costs $18.12 per million input tokens and $81.94 per million output tokens. A blended 3:1 input-to-output mix works out to about $34.08 per million tokens. Cached input is listed at $1.50 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $18.12 | per 1M tokens |
| Output | $81.94 | per 1M tokens |
| Cached input | $1.50 | per 1M tokens |
| Blended 3:1 | $34.08 | estimate per 1M tokens |
| 1M in + 1M out | $100.06 | estimate |
Benchmarks
Scores below are modelled estimates on a 0–100 scale. Use them to compare relative position, not as official vendor results.
| Benchmark | Score | Rank | Percentile | Relative |
|---|---|---|---|---|
| MMLU-Pro | 88.1 | 15 | 98th | |
| GPQA Diamond | 79.8 | 61 | 90th | |
| SWE-bench Verified | 75.3 | 37 | 94th | |
| HumanEval | 92.6 | 14 | 98th | |
| MATH-500 | 93.0 | 11 | 98th | |
| AIME 2025 | 83.1 | 54 | 91th | |
| LiveCodeBench | 78.0 | 64 | 89th | |
| IFEval | 92.2 | 36 | 94th | |
| TruthfulQA | 88.8 | 15 | 98th | |
| HellaSwag | 91.5 | 48 | 92th | |
| MMMU | 82.3 | 29 | 95th | |
| Arena Elo | 1461.0 | 51 | 92th |
What Qwen3-Max Reasoning is good for
Qwen3-Max Reasoning sits in the frontier tier of the index. It is aimed at work where answer quality matters more than unit cost: hard reasoning, multi-step coding, document analysis and multimodal inputs. At $18.12 per million input tokens and a 2,000,000-token context window, it is a premium option, so reserve it for tasks where a cheaper model measurably fails.
Strengths
- Top-tier average benchmark score (85.9), placing it #38 of 600.
- Large 2,000,000-token context window for long documents and agent traces.
- Accepts more than text, including image input where the provider supports it.
- Reasoning-oriented profile, useful for multi-step problems and structured analysis.
- Coding benchmarks are a relative strength within its tier.
Trade-offs
- Proprietary licence: you cannot download the weights or self-host the model.
- Premium input pricing at $18.12 per million tokens.
- Cost can climb quickly on long outputs and high-volume traffic.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| GPT-5.1 | OpenAI | 2000000 | $19.01 | 86.0 |
| Claude Opus 4.5 Vision | Anthropic | 200000 | $17.70 | 85.9 |
| Claude Sonnet 4.5 Turbo | Anthropic | 1000000 | $18.37 | 85.9 |
| Grok 3 Max | xAI | 2000000 | $17.66 | 85.9 |
| Qwen3-Max Max | Alibaba Cloud | 400000 | $18.53 | 85.9 |
| Qwen3-Max Reasoning | Alibaba Cloud | 2000000 | $18.12 | 85.9 |
FAQ
How much does Qwen3-Max Reasoning cost?
Qwen3-Max Reasoning is listed at $18.12 per million input tokens and $81.94 per million output tokens. Cached input is $1.50 per million tokens. A blended 3:1 input-to-output mix is about $34.08 per million tokens.
What is the context window of Qwen3-Max Reasoning?
Qwen3-Max Reasoning has a context window of 2,000,000 tokens and can emit up to 8,192 output tokens per request in this index.
Is Qwen3-Max Reasoning open weights?
No. Qwen3-Max Reasoning is a proprietary model with a proprietary licence, so you cannot download the weights.
How does Qwen3-Max Reasoning score on benchmarks?
Qwen3-Max Reasoning averages 85.9 across the scored benchmarks in this index, ranking #38 of 600 models. Its strongest relative result is TruthfulQA at 88.8.
What is Qwen3-Max Reasoning best for?
Qwen3-Max Reasoning sits in the frontier tier of the index. It is aimed at work where answer quality matters more than unit cost: hard reasoning, multi-step coding, document analysis and multimodal inputs. At $18.12 per million input tokens and a 2,000,000-token context window, it is a premium option, so reserve it for tasks where a cheaper model measurably fails.