Qwen3-Coder: pricing, context window and benchmarks
Qwen3-Coder is a coding model from Alibaba Cloud. It offers a 200,000-token context window, costs $4.94 per million input tokens and $19.51 per million output tokens, and averages 79.5 across the index's benchmark suite.
Pricing
Qwen3-Coder costs $4.94 per million input tokens and $19.51 per million output tokens. A blended 3:1 input-to-output mix works out to about $8.58 per million tokens. Cached input is listed at $0.53 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $4.94 | per 1M tokens |
| Output | $19.51 | per 1M tokens |
| Cached input | $0.53 | per 1M tokens |
| Blended 3:1 | $8.58 | estimate per 1M tokens |
| 1M in + 1M out | $24.45 | 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 | 83.7 | 89 | 85th | |
| GPQA Diamond | 77.9 | 88 | 85th | |
| SWE-bench Verified | 71.0 | 96 | 84th | |
| HumanEval | 83.5 | 146 | 76th | |
| MATH-500 | 86.1 | 92 | 85th | |
| AIME 2025 | 74.2 | 175 | 71th | |
| LiveCodeBench | 69.8 | 196 | 67th | |
| IFEval | 86.6 | 144 | 76th | |
| TruthfulQA | 79.8 | 192 | 68th | |
| HellaSwag | 86.7 | 188 | 69th | |
| MMMU | 75.2 | 159 | 74th | |
| Arena Elo | 1439.3 | 98 | 84th |
What Qwen3-Coder is good for
Qwen3-Coder is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $4.94 per million input tokens, it sits in the mid-range of the index, and the 200,000-token context window is enough for large files and multi-file diffs.
Strengths
- Large 200,000-token context window for long documents and agent traces.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Accepts more than text, including image input where the provider supports it.
- Coding benchmarks are a relative strength within its tier.
Trade-offs
- No standout weakness in the indexed specifications, but real-world behaviour still needs testing.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Command A | Cohere | 200000 | $13.30 | 79.6 |
| Nova Pro Pro | Amazon | 200000 | $13.03 | 79.5 |
| Qwen3-Coder | Alibaba Cloud | 200000 | $4.94 | 79.5 |
| Claude Haiku 4.5 | Anthropic | 1000000 | $13.86 | 79.4 |
| MiniMax-M2 Instruct | MiniMax | 128000 | $5.11 | 79.4 |
| Nova Pro | Amazon | 32768 | $12.75 | 79.4 |
FAQ
How much does Qwen3-Coder cost?
Qwen3-Coder is listed at $4.94 per million input tokens and $19.51 per million output tokens. Cached input is $0.53 per million tokens. A blended 3:1 input-to-output mix is about $8.58 per million tokens.
What is the context window of Qwen3-Coder?
Qwen3-Coder has a context window of 200,000 tokens and can emit up to 8,192 output tokens per request in this index.
Is Qwen3-Coder open weights?
Yes. Qwen3-Coder is listed as an open-weight model under a open weights licence.
How does Qwen3-Coder score on benchmarks?
Qwen3-Coder averages 79.5 across the scored benchmarks in this index, ranking #134 of 600 models. Its strongest relative result is Arena Elo at 1439.3.
What is Qwen3-Coder best for?
Qwen3-Coder is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $4.94 per million input tokens, it sits in the mid-range of the index, and the 200,000-token context window is enough for large files and multi-file diffs.