DeepSeek-Coder 8B: pricing, context window and benchmarks
DeepSeek-Coder 8B is a coding model from DeepSeek. It offers a 32,768-token context window, costs $1.26 per million input tokens and $6.25 per million output tokens, and averages 58.7 across the index's benchmark suite.
Pricing
DeepSeek-Coder 8B costs $1.26 per million input tokens and $6.25 per million output tokens. A blended 3:1 input-to-output mix works out to about $2.51 per million tokens. Cached input is listed at $0.13 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $1.26 | per 1M tokens |
| Output | $6.25 | per 1M tokens |
| Cached input | $0.13 | per 1M tokens |
| Blended 3:1 | $2.51 | estimate per 1M tokens |
| 1M in + 1M out | $7.51 | 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 | 63.2 | 485 | 19th | |
| GPQA Diamond | 54.0 | 481 | 20th | |
| SWE-bench Verified | 44.3 | 472 | 21th | |
| HumanEval | 63.5 | 455 | 24th | |
| MATH-500 | 59.3 | 473 | 21th | |
| AIME 2025 | 51.2 | 447 | 26th | |
| LiveCodeBench | 48.0 | 466 | 22th | |
| IFEval | 68.9 | 474 | 21th | |
| TruthfulQA | 62.3 | 497 | 17th | |
| HellaSwag | 73.0 | 481 | 20th | |
| MMMU | 57.9 | 458 | 24th | |
| Arena Elo | 1260.4 | 474 | 21th |
What DeepSeek-Coder 8B is good for
DeepSeek-Coder 8B 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 $1.26 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
- Input price of $1.26 per million tokens is at or below the index median.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Coding benchmarks are a relative strength within its tier.
Trade-offs
- Average benchmark score is below the index median; test it on your own task.
- Less reliable on open-ended reasoning and complex multi-step instructions.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Codestral 8B | Mistral AI | 4096 | $1.18 | 59.0 |
| Magistral 7B | Mistral AI | 32768 | $1.22 | 58.9 |
| Llama 3.1 Turbo 8B | Together AI | 32768 | $1.16 | 58.8 |
| DeepSeek-Coder 8B | DeepSeek | 32768 | $1.26 | 58.7 |
| Hunyuan-Large 8B | Tencent | 4096 | $1.29 | 58.7 |
| Llama 3.1 8B | Meta AI | 8192 | $1.15 | 58.7 |
FAQ
How much does DeepSeek-Coder 8B cost?
DeepSeek-Coder 8B is listed at $1.26 per million input tokens and $6.25 per million output tokens. Cached input is $0.13 per million tokens. A blended 3:1 input-to-output mix is about $2.51 per million tokens.
What is the context window of DeepSeek-Coder 8B?
DeepSeek-Coder 8B 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 8B open weights?
Yes. DeepSeek-Coder 8B is listed as an open-weight model under a open weights licence.
How does DeepSeek-Coder 8B score on benchmarks?
DeepSeek-Coder 8B averages 58.7 across the scored benchmarks in this index, ranking #468 of 600 models. Its strongest relative result is Arena Elo at 1260.4.
What is DeepSeek-Coder 8B best for?
DeepSeek-Coder 8B 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 $1.26 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.