Claude 4 Sonnet (Reasoning) vs Gemini 3 Pro Preview (high)
Comparing 2 AI models · 6 benchmarks · Anthropic, Google
Composite Indices
Intelligence, Coding, Math
Standard Benchmarks
Academic and industry benchmarks
Benchmark Winners
Claude 4 Sonnet (Reasoning)
- MATH 500
Gemini 3 Pro Preview (high)
- GPQA
- MMLU Pro
- HLE
- LiveCodeBench
- AIME 2025
| Metric | An Claude 4 Sonnet (Reasoning) | Go Gemini 3 Pro Preview (high) |
|---|---|---|
| Pricing Per 1M tokens | ||
| Input Cost | $3.00/1M | $2.00/1M |
| Output Cost | $15.00/1M | $12.00/1M |
| Blended Cost 3:1 input/output ratio | $6.00/1M | $4.50/1M |
| Specifications | ||
| Organization Model creator | Anthropic | |
| Release Date Launch date | May 22, 2025 | Nov 18, 2025 |
| Performance & Speed | ||
| Throughput Output speed | 71.2 tok/s | 138.9 tok/s |
| Time to First Token (TTFT) Initial response delay | 1753ms | 26684ms |
| Latency Time to first answer token | 29860ms | 26684ms |
| Composite Indices | ||
| Intelligence Index Overall reasoning capability | 56.5 | 72.8 |
| Coding Index Programming ability | 45.1 | 62.3 |
| Math Index Mathematical reasoning | 74.3 | 95.7 |
| Standard Benchmarks | ||
| GPQA Graduate-level reasoning | 77.7% | 90.8% |
| MMLU Pro Advanced knowledge | 84.2% | 89.8% |
| HLE Hard language evaluation | 9.6% | 37.2% |
| LiveCodeBench Real-world coding tasks | 65.5% | 91.7% |
| MATH 500 Mathematical problems | 99.1% | — |
| AIME 2025 Advanced math competition | 74.3% | 95.7% |
| AIME (Original) Math olympiad problems | 77.3% | — |
| SciCode Scientific code generation | 40.0% | 56.1% |
| LCR Code review capability | 64.7% | 70.7% |
| IFBench Instruction-following | 54.7% | 70.4% |
| TAU-bench v2 Tool use & agentic tasks | 64.6% | 87.1% |
| TerminalBench Hard CLI command generation | 29.8% | 39.0% |
Key Takeaways
Gemini 3 Pro Preview (high) offers the best value at $2.00/1M, making it ideal for high-volume applications and cost-conscious projects.
Gemini 3 Pro Preview (high) leads in reasoning capabilities with a 90.8% GPQA score, excelling at complex analytical tasks and problem-solving.
Gemini 3 Pro Preview (high) achieves a 62.3 coding index, making it the top choice for software development and code generation tasks.
All models support context windows of ∞+ tokens, suitable for processing lengthy documents and maintaining extended conversations.
When to Choose Each Model
Claude 4 Sonnet (Reasoning)
- General-purpose AI
- Versatile applications
Gemini 3 Pro Preview (high)
- Cost-sensitive applications
- High-volume processing
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
Try Models for Free
Try Claude 4 Sonnet (Reasoning) for FREE
No credit card or account required.
Try Gemini 3 Pro Preview (high) for FREE
No credit card or account required.
Cost Calculator
Claude 4 Sonnet (Reasoning) Gemini 3 Pro Preview (high) Costs are estimates based on API pricing. Actual costs may vary based on caching, batch processing, and volume discounts.
Recommended Comparisons
