Provider Snapshot
25
54.19
1040.80
May 20, 2026
Key Takeaways
25 bedrock models are actively benchmarked with 20213 total measurements across 18027 benchmark runs.
nova-micro leads the fleet with 122.00 tokens/second, while nova-pro delivers 86.50 tok/s.
Performance varies by 41.0% across the bedrock model lineup, indicating diverse optimization strategies for different use cases.
Avg time to first token across the fleet is 1040.80 ms, showing moderate responsiveness for interactive applications.
The bedrock model fleet shows varied performance characteristics (60.7% variation coefficient), reflecting diverse model architectures.
Fastest Models
| Provider | Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|---|
| bedrock | nova-micro | 122.00 | 69.70 | 149.00 | 270.00 |
| bedrock | llama-4-maverick | 107.00 | 46.70 | 134.00 | 320.00 |
| bedrock | nova-lite | 97.40 | 56.60 | 123.00 | 310.00 |
| bedrock | llama-4-scout | 94.50 | 1.56 | 124.00 | 330.00 |
| bedrock | llama-3.1-8b | 91.30 | 3.06 | 109.00 | 720.00 |
| bedrock | nova-pro | 86.50 | 19.80 | 119.00 | 370.00 |
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 | 97.40 | 56.60 | 123.00 | 310.00 |
| bedrock | nova-micro | 122.00 | 69.70 | 149.00 | 270.00 |
| bedrock | nova-pro | 86.50 | 19.80 | 119.00 | 370.00 |
| bedrock | llama-3.1-405b | 17.30 | 7.89 | 19.70 | 3720.00 |
| bedrock | llama-3.1-70b | 24.70 | 19.90 | 25.80 | 610.00 |
| bedrock | llama-3.1-8b | 91.30 | 3.06 | 109.00 | 720.00 |
| bedrock | llama-3-70b | 37.50 | 20.50 | 41.60 | 270.00 |
| bedrock | llama-3-8b | 77.70 | 54.30 | 83.20 | 220.00 |
| bedrock | mistral-7b | 80.30 | 57.90 | 86.80 | 200.00 |
| bedrock | mistral-large | 40.70 | 5.26 | 45.90 | 450.00 |
| bedrock | mistral-small | 56.60 | 43.10 | 59.40 | 210.00 |
| bedrock | mixtral-8x7b | 76.70 | 20.70 | 82.80 | 230.00 |
| bedrock | claude-3-5-haiku | 29.70 | 1.02 | 36.90 | 760.00 |
| bedrock | claude-3-7-sonnet | 32.60 | 3.97 | 41.10 | 870.00 |
| bedrock | claude-haiku-4.5 | 43.80 | 6.60 | 59.10 | 930.00 |
| bedrock | claude-opus-4.1 | 9.15 | 4.87 | 15.30 | 3140.00 |
| bedrock | claude-opus-4.5 | 19.40 | 3.90 | 23.00 | 1780.00 |
| bedrock | claude-opus-4.6 | 20.90 | 8.48 | 24.90 | 1690.00 |
| bedrock | claude-opus-4.7 | 19.40 | 1.09 | 38.80 | 4480.00 |
| bedrock | claude-sonnet-4.5 | 21.20 | 4.29 | 27.70 | 1570.00 |
| bedrock | claude-sonnet-4.6 | 24.50 | 1.15 | 33.00 | 1850.00 |
| bedrock | llama-3.2-90b | 45.90 | 31.60 | 49.50 | 410.00 |
| bedrock | llama-3.3-70b | 78.10 | 16.80 | 113.00 | 310.00 |
| bedrock | llama-4-maverick | 107.00 | 46.70 | 134.00 | 320.00 |
| bedrock | llama-4-scout | 94.50 | 1.56 | 124.00 | 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 20213 individual prompts measured across 18027 monitoring runs over the past month.