> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scaledown.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Head-to-head

> ScaleDown versus each provider's lineup on accuracy, cost and latency.

Every figure is ScaleDown measured against that provider's **average across its models** on the same task. Accuracy is the gain in score; cost and latency are how many times cheaper or faster ScaleDown is. Datasets and metrics are in [Methodology](/benchmarks/methodology).

## OpenAI

**Models compared:** GPT-5.5, GPT-5.4 Mini, GPT-5.4 Nano

<CardGroup cols={3}>
  <Card title="8.72% higher accuracy" icon="bullseye">than the OpenAI avg</Card>
  <Card title="89× cheaper" icon="dollar-sign">than the OpenAI avg</Card>
  <Card title="2.4× faster" icon="bolt">than the OpenAI avg</Card>
</CardGroup>

| Task | Accuracy | Cost | Latency |
| - | - | - | - |
| Classify | +5.40% | 1,810× cheaper | 2.6× faster |
| Extract | +14.40% | 35× cheaper | 3.3× faster |
| Summarize | +8.83% | 41× cheaper | 2.7× faster |
| Compress | +6.23% | 138× cheaper | 1.8× faster |

[Full ScaleDown vs OpenAI charts and cost tables](https://scaledown.ai/benchmarks/scaledown-vs-openai)

## Anthropic

**Models compared:** Fable 5, Opus 4.8, Sonnet 4.6, Haiku 4.5

<CardGroup cols={3}>
  <Card title="8.0% higher accuracy" icon="bullseye">than the Anthropic avg</Card>
  <Card title="161× cheaper" icon="dollar-sign">than the Anthropic avg</Card>
  <Card title="3.8× faster" icon="bolt">than the Anthropic avg</Card>
</CardGroup>

| Task | Accuracy | Cost | Latency |
| - | - | - | - |
| Classify | +2.29% | 5,250× cheaper | 3.6× faster |
| Extract | +13.33% | 91× cheaper | 3.2× faster |
| Summarize | +8.22% | 217× cheaper | 4.4× faster |

[Full ScaleDown vs Anthropic charts and cost tables](https://scaledown.ai/benchmarks/scaledown-vs-anthropic)

## Gemini

**Models compared:** 3.1 Pro, 3.5 Flash, 3.1 Flash-Lite

<CardGroup cols={3}>
  <Card title="9.0% higher accuracy" icon="bullseye">than the Gemini avg</Card>
  <Card title="29× cheaper" icon="dollar-sign">than the Gemini avg</Card>
  <Card title="8.3× faster" icon="bolt">than the Gemini avg</Card>
</CardGroup>

| Task | Accuracy | Cost | Latency |
| - | - | - | - |
| Classify | +2.10% | 2,437× cheaper | 13.8× faster |
| Extract | +10.34% | 26× cheaper | 13.4× faster |
| Summarize | +14.70% | 28× cheaper | 2.6× faster |

[Full ScaleDown vs Gemini charts and cost tables](https://scaledown.ai/benchmarks/scaledown-vs-gemini)

## Grok

**Models compared:** Grok 4.3, Grok 4.2

<CardGroup cols={3}>
  <Card title="7.93% higher accuracy" icon="bullseye">than the Grok avg</Card>
  <Card title="24× cheaper" icon="dollar-sign">than the Grok avg</Card>
  <Card title="2.2× faster" icon="bolt">than the Grok avg</Card>
</CardGroup>

| Task | Accuracy | Cost | Latency |
| - | - | - | - |
| Classify | +1.45% | 1,161× cheaper | 2.3× faster |
| Extract | +12.30% | 22× cheaper | 2.3× faster |
| Summarize | +10.05% | 25× cheaper | 2.1× faster |

[Full ScaleDown vs Grok charts and cost tables](https://scaledown.ai/benchmarks/scaledown-vs-grok)

## DeepSeek

**Models compared:** DeepSeek V4 Pro, DeepSeek V4 Flash

<CardGroup cols={3}>
  <Card title="19.5% higher accuracy" icon="bullseye">than the DeepSeek avg</Card>
  <Card title="9× cheaper" icon="dollar-sign">than the DeepSeek avg</Card>
  <Card title="4.1× faster" icon="bolt">than the DeepSeek avg</Card>
</CardGroup>

| Task | Accuracy | Cost | Latency |
| - | - | - | - |
| Classify | +20.75% | 113× cheaper | 3.7× faster |
| Extract | +31.03% | 9× cheaper | 3.9× faster |
| Summarize | +6.73% | 10× cheaper | 4.4× faster |

[Full ScaleDown vs DeepSeek charts and cost tables](https://scaledown.ai/benchmarks/scaledown-vs-deepseek)

## Qwen

**Models compared:** Qwen3.8 Max, Qwen3.8 27B, Qwen3.5 4B

<CardGroup cols={3}>
  <Card title="16.71% higher accuracy" icon="bullseye">than the Qwen avg</Card>
  <Card title="24× cheaper" icon="dollar-sign">than the Qwen avg</Card>
  <Card title="13.9× faster" icon="bolt">than the Qwen avg</Card>
</CardGroup>

| Task | Accuracy | Cost | Latency |
| - | - | - | - |
| Classify | +9.12% | 761× cheaper | 7.2× faster |
| Extract | +31.98% | 18× cheaper | 26.1× faster |
| Summarize | +9.04% | 28× cheaper | 9.2× faster |

[Full ScaleDown vs Qwen charts and cost tables](https://scaledown.ai/benchmarks/scaledown-vs-qwen)

## GLiNER

**Model compared:** GLiNER2.5-multi-v1, a local open-source model for classification and extraction.

GLiNER is the one comparison where ScaleDown is not the faster or cheaper option. GLiNER runs locally, so inference is free, and it is faster, while ScaleDown is more accurate.

| Task | Accuracy (ScaleDown vs GLiNER) | GLiNER latency | Cost |
| - | - | - | - |
| Classify | +15.5 pp for ScaleDown | 22 ms | GLiNER free (local inference) |
| Extract | +0.2 pp for ScaleDown | 480 ms | GLiNER free (local inference) |

Averaged across both tasks, ScaleDown's accuracy advantage is +10.5%.

[Full ScaleDown vs GLiNER charts](https://scaledown.ai/benchmarks/scaledown-vs-gliner)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.