Provider Snapshot
29
62.26
0.57
Oct 29, 2025
Key Takeaways
29 bedrock models are actively benchmarked with 3723 total measurements across 3430 benchmark runs.
nova-micro leads the fleet with 138.00 tokens/second, while llama-3.3-70b delivers 100.00 tok/s.
Performance varies by 38.0% across the bedrock model lineup, indicating diverse optimization strategies for different use cases.
Median time to first token across the fleet is 0.57 ms, showing excellent responsiveness for interactive applications.
The bedrock model fleet shows varied performance characteristics (53.4% variation coefficient), reflecting diverse model architectures.
Fastest Models
| Provider | Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|---|
| bedrock | nova-micro | 138.00 | 114.00 | 163.00 | 0.18 |
| bedrock | llama-4-maverick | 116.00 | 77.90 | 152.00 | 0.23 |
| bedrock | nova-lite | 109.00 | 3.93 | 150.00 | 0.22 |
| bedrock | llama-4-scout | 106.00 | 72.60 | 126.00 | 0.25 |
| bedrock | llama-3.1-8b | 101.00 | 35.80 | 113.00 | 0.32 |
| bedrock | llama-3.3-70b | 100.00 | 22.20 | 128.00 | 0.23 |
All Models
Complete list of all bedrock models tracked in the benchmark system. Click any model name to view detailed performance data.
| Provider | Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|---|
| bedrock | claude-3-haiku | 62.20 | 18.10 | 85.50 | 0.26 |
| bedrock | claude-3-5-haiku | 31.40 | 6.93 | 37.60 | 0.94 |
| bedrock | claude-3-sonnet | 44.50 | 15.30 | 55.60 | 0.28 |
| bedrock | claude-3-5-sonnet | 33.60 | 16.70 | 44.50 | 0.97 |
| bedrock | claude-3-7-sonnet | 28.60 | 19.00 | 39.00 | 1.21 |
| bedrock | claude-3-opus | 18.90 | 13.00 | 21.90 | 0.84 |
| bedrock | titan-text-lite | 26.50 | 20.40 | 30.60 | 2.42 |
| bedrock | llama-3-8b | 82.40 | 68.20 | 86.80 | 0.16 |
| bedrock | llama-3-70b | 39.90 | 33.00 | 42.10 | 0.22 |
| bedrock | llama-3.1-8b | 101.00 | 35.80 | 113.00 | 0.32 |
| bedrock | llama-3.1-70b | 25.40 | 21.00 | 26.20 | 0.54 |
| bedrock | llama-3.1-405b | 18.30 | 1.59 | 20.10 | 1.68 |
| bedrock | llama-3.2-1b | 62.50 | 50.30 | 70.90 | 0.30 |
| bedrock | llama-3.2-3b | 47.80 | 35.80 | 58.60 | 0.32 |
| bedrock | llama-3.2-11b | 92.50 | 23.10 | 117.00 | 0.31 |
| bedrock | llama-3.2-90b | 45.10 | 34.70 | 49.20 | 0.41 |
| bedrock | mistral-7b | 81.00 | 29.70 | 91.90 | 0.14 |
| bedrock | mixtral-8x7b | 78.80 | 22.00 | 88.30 | 0.17 |
| bedrock | mistral-small | 60.00 | 52.50 | 62.70 | 0.15 |
| bedrock | mistral-large | 43.20 | 8.39 | 47.40 | 0.29 |
| bedrock | command-r | 75.10 | 28.90 | 79.50 | 0.13 |
| bedrock | command-r-plus | 35.90 | 24.80 | 37.10 | 0.16 |
| bedrock | nova-pro | 84.90 | 4.47 | 127.00 | 0.30 |
| bedrock | nova-lite | 109.00 | 3.93 | 150.00 | 0.22 |
| bedrock | nova-micro | 138.00 | 114.00 | 163.00 | 0.18 |
| bedrock | llama-3.3-70b | 100.00 | 22.20 | 128.00 | 0.23 |
| bedrock | llama-4-maverick | 116.00 | 77.90 | 152.00 | 0.23 |
| bedrock | llama-4-scout | 106.00 | 72.60 | 126.00 | 0.25 |
| bedrock | claude-sonnet-4.5 | 17.10 | 5.25 | 21.70 | 2.87 |
Featured Models
Frequently Asked Questions
Based on recent tests, nova-micro shows the highest average throughput among tracked bedrock models.
This provider summary aggregates 3723 individual prompts measured across 3430 monitoring runs over the past month.