Sonar Turbo: pricing, context window and benchmarks
Sonar Turbo is a multimodal model from Perplexity. It offers a 65,536-token context window, costs $10.71 per million input tokens and $43.83 per million output tokens, and averages 75.7 across the index's benchmark suite.
Pricing
Sonar Turbo costs $10.71 per million input tokens and $43.83 per million output tokens. A blended 3:1 input-to-output mix works out to about $18.99 per million tokens. Cached input is listed at $1.51 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $10.71 | per 1M tokens |
| Output | $43.83 | per 1M tokens |
| Cached input | $1.51 | per 1M tokens |
| Blended 3:1 | $18.99 | estimate per 1M tokens |
| 1M in + 1M out | $54.54 | 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 | 80.4 | 172 | 71th | |
| GPQA Diamond | 73.1 | 197 | 67th | |
| SWE-bench Verified | 66.3 | 188 | 69th | |
| HumanEval | 78.7 | 272 | 55th | |
| MATH-500 | 78.8 | 243 | 60th | |
| AIME 2025 | 71.0 | 237 | 61th | |
| LiveCodeBench | 67.9 | 228 | 62th | |
| IFEval | 82.7 | 254 | 58th | |
| TruthfulQA | 79.1 | 213 | 65th | |
| HellaSwag | 83.6 | 301 | 50th | |
| MMMU | 71.2 | 268 | 55th | |
| Arena Elo | 1387.3 | 258 | 57th |
What Sonar Turbo is good for
Sonar Turbo is a general-purpose model in the index. It handles text generation, summarization and question answering, and it can be a sensible default when you need a balance of quality, price and context. At $10.71 per million input tokens with a 65,536-token window, check its benchmark profile against the specific task before committing.
Strengths
- Accepts more than text, including image input where the provider supports it.
Trade-offs
- Proprietary licence: you cannot download the weights or self-host the model.
- Premium input pricing at $10.71 per million tokens.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Hunyuan-Turbo | Tencent | 128000 | $10.55 | 75.8 |
| Llama 3 Lambda Instruct | Lambda | 32768 | $3.92 | 75.8 |
| Llama 3 Anyscale Instruct | Anyscale | 200000 | $3.91 | 75.7 |
| Llama 3 Replicate | Replicate | 200000 | $3.75 | 75.7 |
| Llama 3.1 Instruct | Meta AI | 200000 | $4.05 | 75.7 |
| Sonar Turbo | Perplexity | 65536 | $10.71 | 75.7 |
FAQ
How much does Sonar Turbo cost?
Sonar Turbo is listed at $10.71 per million input tokens and $43.83 per million output tokens. Cached input is $1.51 per million tokens. A blended 3:1 input-to-output mix is about $18.99 per million tokens.
What is the context window of Sonar Turbo?
Sonar Turbo has a context window of 65,536 tokens and can emit up to 8,192 output tokens per request in this index.
Is Sonar Turbo open weights?
No. Sonar Turbo is a proprietary model with a proprietary licence, so you cannot download the weights.
How does Sonar Turbo score on benchmarks?
Sonar Turbo averages 75.7 across the scored benchmarks in this index, ranking #233 of 600 models. Its strongest relative result is Arena Elo at 1387.3.
What is Sonar Turbo best for?
Sonar Turbo is a general-purpose model in the index. It handles text generation, summarization and question answering, and it can be a sensible default when you need a balance of quality, price and context. At $10.71 per million input tokens with a 65,536-token window, check its benchmark profile against the specific task before committing.