Updated March 25, 2026· Based on independent benchmark data
Anthropic: Claude Sonnet 4 leads in intelligence with a score of 51.7 vs 48.4. For speed, Google: Gemini 3 Pro Preview wins at 116 tok/s vs 65 tok/s.
| Metric | Anthropic: Claude Sonnet 4 | Google: Gemini 3 Pro Preview |
|---|---|---|
| Intelligence Score | 51.7 | 48.4 |
| Coding Score | 50.9 | 46.5 |
| Math Score | N/A | 95.7 |
| Speed (tok/s) | 65 tok/s | 116 tok/s |
| Latency (TTFT) | 38.68s | 25.53s |
| Input Price / 1M tokens | $3.00 | $2.00 |
| Output Price / 1M tokens | $15 | $12 |
| Context Window | 200K |
Anthropic: Claude Sonnet 4 outperforms Google: Gemini 3 Pro Preview on the Artificial Analysis intelligence index with a score of 51.7 compared to 48.4. For coding tasks, Anthropic: Claude Sonnet 4 has the edge with a coding score of 50.9 vs 46.5.
Google: Gemini 3 Pro Preview generates output significantly faster at 116 tok/s compared to Anthropic: Claude Sonnet 4's 65 tok/s, making it 1.8x faster for streaming responses. Time to first token is 25.53s for Google: Gemini 3 Pro Preview vs 38.68s for Anthropic: Claude Sonnet 4, which affects perceived responsiveness in interactive applications.
Google: Gemini 3 Pro Preview is more affordable at $2.00/1M input tokens ($12/1M output), while Anthropic: Claude Sonnet 4 costs $3.00/1M input ($15/1M output). For a typical workload of 100 requests per day at 2,000 tokens each, Anthropic: Claude Sonnet 4 would cost approximately $18.00/month vs $12.00/month for Google: Gemini 3 Pro Preview in input costs alone.
Google: Gemini 3 Pro Preview offers a larger context window at 1.0M tokens compared to Anthropic: Claude Sonnet 4's 200K. This means Google: Gemini 3 Pro Preview can process roughly 524 pages of text in a single request vs 100 pages for Anthropic: Claude Sonnet 4. For output length, Google: Gemini 3 Pro Preview can generate up to 66K tokens per response vs 64K for Anthropic: Claude Sonnet 4.
Choose Anthropic: Claude Sonnet 4 when you need higher intelligence (51.7), stronger coding performance (50.9). Choose Google: Gemini 3 Pro Preview when you need faster output (116 tok/s), larger context window (1.0M).
Anthropic: Claude Sonnet 4 scores higher on coding benchmarks (50.9 vs 46.5), making it the better choice for programming tasks.
Google: Gemini 3 Pro Preview is cheaper at $2.00/1M input tokens vs $3.00/1M for Anthropic: Claude Sonnet 4.
Google: Gemini 3 Pro Preview is faster, producing output at 116 tok/s compared to Anthropic: Claude Sonnet 4's 65 tok/s.
Yes, Anthropic: Claude Sonnet 4 supports image input. Google: Gemini 3 Pro Preview also supports images.
Benchmark data by Artificial Analysis
Data last synced: March 25, 2026
| 1.0M |
| Max Output Tokens | 64K | 66K |
| Input Modalities | Image + Text + File | Text + Image + File + Audio + Video |
| Output Modalities | Text | Text |
| Free Tier | No | No |
Google: Gemini 3 Pro Preview has a larger context window at 1.0M compared to Anthropic: Claude Sonnet 4's 200K.
It depends on your priorities. Anthropic: Claude Sonnet 4 scores higher on intelligence (51.7), but Google: Gemini 3 Pro Preview may be better for specific use cases like budget-conscious projects or speed-critical applications.