DeepSeek-Coder Instruct: pricing, context window and benchmarks
DeepSeek-Coder Instruct is a coding model from DeepSeek. It offers a 32,768-token context window, costs $4.34 per million input tokens and $20.00 per million output tokens, and averages 77.7 across the index's benchmark suite.
Pricing
DeepSeek-Coder Instruct costs $4.34 per million input tokens and $20.00 per million output tokens. A blended 3:1 input-to-output mix works out to about $8.25 per million tokens. Cached input is listed at $0.55 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $4.34 | per 1M tokens |
| Output | $20.00 | per 1M tokens |
| Cached input | $0.55 | per 1M tokens |
| Blended 3:1 | $8.25 | estimate per 1M tokens |
| 1M in + 1M out | $24.34 | 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 | 81.5 | 134 | 78th | |
| GPQA Diamond | 70.6 | 243 | 60th | |
| SWE-bench Verified | 67.2 | 165 | 73th | |
| HumanEval | 81.6 | 192 | 68th | |
| MATH-500 | 81.1 | 189 | 69th | |
| AIME 2025 | 73.5 | 196 | 67th | |
| LiveCodeBench | 71.7 | 154 | 74th | |
| IFEval | 84.7 | 197 | 67th | |
| TruthfulQA | 80.6 | 165 | 73th | |
| HellaSwag | 86.8 | 183 | 70th | |
| MMMU | 75.5 | 150 | 75th | |
| Arena Elo | 1392.8 | 237 | 61th |
What DeepSeek-Coder Instruct is good for
DeepSeek-Coder Instruct 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 $4.34 per million input tokens, it sits in the mid-range of the index, and the 32,768-token context window is enough for large files and multi-file diffs.
Strengths
- 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 |
|---|---|---|---|---|
| MiniMax-Text Chat | MiniMax | 128000 | $4.26 | 77.8 |
| Sonar Max | Perplexity | 32768 | $11.60 | 77.8 |
| DeepSeek-Coder Instruct | DeepSeek | 32768 | $4.34 | 77.7 |
| GPT-5.1 Mini | OpenAI | 128000 | $12.19 | 77.7 |
| MiniMax-Text Instruct | MiniMax | 128000 | $4.59 | 77.7 |
| o4 Mini | OpenAI | 200000 | $11.92 | 77.7 |
FAQ
How much does DeepSeek-Coder Instruct cost?
DeepSeek-Coder Instruct is listed at $4.34 per million input tokens and $20.00 per million output tokens. Cached input is $0.55 per million tokens. A blended 3:1 input-to-output mix is about $8.25 per million tokens.
What is the context window of DeepSeek-Coder Instruct?
DeepSeek-Coder Instruct has a context window of 32,768 tokens and can emit up to 32,768 output tokens per request in this index.
Is DeepSeek-Coder Instruct open weights?
Yes. DeepSeek-Coder Instruct is listed as an open-weight model under a open weights licence.
How does DeepSeek-Coder Instruct score on benchmarks?
DeepSeek-Coder Instruct averages 77.7 across the scored benchmarks in this index, ranking #179 of 600 models. Its strongest relative result is Arena Elo at 1392.8.
What is DeepSeek-Coder Instruct best for?
DeepSeek-Coder Instruct 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 $4.34 per million input tokens, it sits in the mid-range of the index, and the 32,768-token context window is enough for large files and multi-file diffs.