Provider Snapshot
24
55.28
957.92
Jun 24, 2026
Key Takeaways
24 bedrock models are actively benchmarked with 33334 total measurements across 30475 benchmark runs.
nova-micro leads the fleet with 118.00 tokens/second, while nova-pro delivers 88.10 tok/s.
Performance varies by 33.9% across the bedrock model lineup, indicating diverse optimization strategies for different use cases.
Avg time to first token across the fleet is 957.92 ms, showing good responsiveness for interactive applications.
The bedrock model fleet shows varied performance characteristics (59.0% variation coefficient), reflecting diverse model architectures.
Fastest Models
| Provider | Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|---|
| bedrock | nova-micro | 118.00 | 2.66 | 148.00 | 300.00 |
| bedrock | llama-4-maverick | 104.00 | 28.90 | 136.00 | 300.00 |
| bedrock | llama-3.1-8b | 95.10 | 3.79 | 111.00 | 450.00 |
| bedrock | llama-4-scout | 94.10 | 1.90 | 123.00 | 330.00 |
| bedrock | nova-lite | 93.30 | 5.96 | 131.00 | 330.00 |
| bedrock | nova-pro | 88.10 | 6.72 | 125.00 | 390.00 |
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 | nova-lite | 93.30 | 5.96 | 131.00 | 330.00 |
| bedrock | nova-micro | 118.00 | 2.66 | 148.00 | 300.00 |
| bedrock | nova-pro | 88.10 | 6.72 | 125.00 | 390.00 |
| bedrock | llama-3.1-405b | 17.10 | 6.85 | 19.40 | 3730.00 |
| bedrock | llama-3.1-70b | 24.80 | 1.55 | 25.90 | 630.00 |
| bedrock | llama-3.1-8b | 95.10 | 3.79 | 111.00 | 450.00 |
| bedrock | llama-3-70b | 35.50 | 4.82 | 41.70 | 330.00 |
| bedrock | llama-3-8b | 78.60 | 11.40 | 84.70 | 210.00 |
| bedrock | mistral-7b | 82.40 | 1.05 | 88.60 | 220.00 |
| bedrock | mistral-large | 42.60 | 3.89 | 46.60 | 280.00 |
| bedrock | mistral-small | 57.00 | 29.40 | 60.30 | 200.00 |
| bedrock | mixtral-8x7b | 75.00 | 1.05 | 81.30 | 250.00 |
| bedrock | claude-3-5-haiku | 29.90 | 1.02 | 37.40 | 1300.00 |
| bedrock | claude-haiku-4.5 | 43.60 | 1.02 | 63.00 | 960.00 |
| bedrock | claude-opus-4.1 | 9.12 | 1.57 | 15.60 | 3200.00 |
| bedrock | claude-opus-4.5 | 20.10 | 8.21 | 24.40 | 1640.00 |
| bedrock | claude-opus-4.6 | 20.70 | 2.11 | 25.60 | 1770.00 |
| bedrock | claude-opus-4.7 | 24.80 | 1.02 | 37.90 | 2520.00 |
| bedrock | claude-sonnet-4.5 | 21.60 | 1.62 | 28.00 | 1500.00 |
| bedrock | claude-sonnet-4.6 | 25.30 | 1.23 | 33.10 | 1270.00 |
| bedrock | llama-3.2-90b | 46.40 | 1.04 | 49.70 | 390.00 |
| bedrock | llama-3.3-70b | 79.70 | 1.03 | 118.00 | 490.00 |
| bedrock | llama-4-maverick | 104.00 | 28.90 | 136.00 | 300.00 |
| bedrock | llama-4-scout | 94.10 | 1.90 | 123.00 | 330.00 |
Featured Models
Frequently Asked Questions
Based on recent tests, nova-micro shows the highest average throughput among tracked bedrock models.
This provider summary aggregates 33334 individual prompts measured across 30475 monitoring runs over the past month.