Llama 3.1 8B 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Healthy | 100%◆ (best) | 99.27% | 99.06% | 98.76% | 514ms | 2.90s | 16 tps | 32K | $0.152 / $0.287 | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.97%◆ (best) | 185ms◆ (best) | 1.75s◆ (best) | 137 tps | 131K | $0.220 / $0.220 | |
Groqgroq | Healthy | 99.89% | 99.66% | 99.69% | 99.64% | 425ms | 5.00s | 185 tps◆ (best) | 131K | $0.050 / $0.080 |
| Healthy | 99.86% | 99.91% | 99.07% | 99.19% | 719ms | 3.20s | 27 tps | 131K | $0.020 / $0.040◆ (best) | |
| Healthy | 99.82% | 98.92% | 99.14% | 98.66% | 545ms | 3.80s | 87 tps | 16K | $0.020 / $0.050◆ (best) |
- Cloudflare100%◆ (best)Healthy24h 99.27%
- 7d
- 99.06%
- 30d
- 98.76%
- p50
- 514ms
- p99
- 2.90s
- Throughput
- 16 tps
- Context
- 32K
- $/Mtok
- $0.152 / $0.287
cloudflare/fp8fp8 - CoreWeave100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.97%◆ (best)
- p50
- 185ms◆ (best)
- p99
- 1.75s◆ (best)
- Throughput
- 137 tps
- Context
- 131K
- $/Mtok
- $0.220 / $0.220
coreweave/bf16bf16 - Groq99.89%Healthy24h 99.66%
- 7d
- 99.69%
- 30d
- 99.64%
- p50
- 425ms
- p99
- 5.00s
- Throughput
- 185 tps◆ (best)
- Context
- 131K
- $/Mtok
- $0.050 / $0.080
groq - DeepInfra99.86%Healthy24h 99.91%
- 7d
- 99.07%
- 30d
- 99.19%
- p50
- 719ms
- p99
- 3.20s
- Throughput
- 27 tps
- Context
- 131K
- $/Mtok
- $0.020 / $0.040◆ (best)
deepinfra/fp8fp8 - Novita99.82%Healthy24h 98.92%
- 7d
- 99.14%
- 30d
- 98.66%
- p50
- 545ms
- p99
- 3.80s
- Throughput
- 87 tps
- Context
- 16K
- $/Mtok
- $0.020 / $0.050◆ (best)
novita/fp8fp8
Measured 3m 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.
Throughput
Median tokens per second.