Zephyr 8B: pricing, context window and benchmarks
Zephyr 8B is a small open-weight model from Hugging Face. It offers a 8,192-token context window, costs $0.22 per million input tokens and $0.84 per million output tokens, and averages 43.0 across the index's benchmark suite.
Pricing
Zephyr 8B costs $0.22 per million input tokens and $0.84 per million output tokens. A blended 3:1 input-to-output mix works out to about $0.38 per million tokens. Cached input is listed at $0.03 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $0.22 | per 1M tokens |
| Output | $0.84 | per 1M tokens |
| Cached input | $0.03 | per 1M tokens |
| Blended 3:1 | $0.38 | estimate per 1M tokens |
| 1M in + 1M out | $1.06 | 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 | 51.8 | 586 | 2th | |
| GPQA Diamond | 39.2 | 587 | 2th | |
| SWE-bench Verified | 26.1 | 594 | 1th | |
| HumanEval | 46.3 | 588 | 2th | |
| MATH-500 | 39.2 | 593 | 1th | |
| AIME 2025 | 26.6 | 590 | 2th | |
| LiveCodeBench | 28.8 | 586 | 2th | |
| IFEval | 56.1 | 590 | 2th | |
| TruthfulQA | 50.9 | 588 | 2th | |
| HellaSwag | 64.8 | 586 | 2th | |
| MMMU | 43.7 | 584 | 3th | |
| Arena Elo | 1133.1 | 591 | 2th |
What Zephyr 8B is good for
Zephyr 8B is a small model. That makes it a good fit for classification, extraction, routing, summarization and other high-volume jobs where latency and cost dominate. It is not the right pick for open-ended frontier reasoning, but at $0.22 per million input tokens it can handle a large share of production traffic.
Strengths
- Input price of $0.22 per million tokens is at or below the index median.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
Trade-offs
- Context window of 8,192 tokens is small for long-document work.
- 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 |
|---|---|---|---|---|
| GPT-NeoX 8B | EleutherAI | 16384 | $0.25 | 43.7 |
| Arctic 1.5B | Snowflake | 32768 | $0.15 | 43.2 |
| Zephyr 8B | Hugging Face | 8192 | $0.22 | 43.0 |
| Molmo 1.5B | Allen Institute for AI | 8192 | $0.21 | 42.8 |
| Zephyr 7B | Hugging Face | 32768 | $0.23 | 42.5 |
| GPT-NeoX 7B | EleutherAI | 8192 | $0.20 | 42.4 |
FAQ
How much does Zephyr 8B cost?
Zephyr 8B is listed at $0.22 per million input tokens and $0.84 per million output tokens. Cached input is $0.032 per million tokens. A blended 3:1 input-to-output mix is about $0.38 per million tokens.
What is the context window of Zephyr 8B?
Zephyr 8B has a context window of 8,192 tokens and can emit up to 4,096 output tokens per request in this index.
Is Zephyr 8B open weights?
Yes. Zephyr 8B is listed as an open-weight model under a open weights licence.
How does Zephyr 8B score on benchmarks?
Zephyr 8B averages 43.0 across the scored benchmarks in this index, ranking #589 of 600 models. Its strongest relative result is Arena Elo at 1133.1.
What is Zephyr 8B best for?
Zephyr 8B is a small model. That makes it a good fit for classification, extraction, routing, summarization and other high-volume jobs where latency and cost dominate. It is not the right pick for open-ended frontier reasoning, but at $0.22 per million input tokens it can handle a large share of production traffic.