Llama 3.1 70B Instruct: providers compared
Every endpoint serving this model, side by side. ◆ marks the best value in each column — per model, never fleet-wide.
Current state
| State | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Amazon Bedrockamazon-bedrock | Healthy | 100%◆ (best) | 100% | 99.97%◆ (best) | 99.40% | 350ms | 1.96s | 71 tps◆ (best) | 131K | $0.720 / $0.720 |
| Healthy | 100%◆ (best) | 99.47% | 96.99% | 94.83% | 310ms◆ (best) | 1.61s◆ (best) | 37 tps | 131K | $0.400 / $0.400◆ (best) | |
| No data | — | — | — | 100%◆ (best) | — | — | — | 131K | $0.800 / $0.800 |
- Amazon Bedrock100%◆ (best)Healthy24h 100%
- 7d
- 99.97%◆ (best)
- 30d
- 99.40%
- p50
- 350ms
- p99
- 1.96s
- Throughput
- 71 tps◆ (best)
- Context
- 131K
- $/Mtok
- $0.720 / $0.720
amazon-bedrock - DeepInfra100%◆ (best)Healthy24h 99.47%
- 7d
- 96.99%
- 30d
- 94.83%
- p50
- 310ms◆ (best)
- p99
- 1.61s◆ (best)
- Throughput
- 37 tps
- Context
- 131K
- $/Mtok
- $0.400 / $0.400◆ (best)
deepinfra/turbofp8 - No data24h —
- 7d
- —
- 30d
- 100%◆ (best)
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 131K
- $/Mtok
- $0.800 / $0.800
coreweave/bf16bf16
Measured 2m ago. Latency and throughput are OpenRouter's rolling 30-minute windows. ◆ marks the best value in each column among these endpoints.
Price vs latency
One dot per endpoint at the latest round · further left = faster, lower = cheaper · hover a dot for its tag
History
Availability
Axis starts at 90% so small differences stay visible; it expands to 0 when an endpoint falls below.
Time to first token
Logarithmic axis — the fleet spans two orders of magnitude. P50 is OpenRouter's rolling 30-minute window, averaged over the hour.
Throughput
Median tokens per second.