Qwen3 32B: 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% | 98.69% | 97.46% | 1.27s | 14.3s | 40 tps◆ (best) | 131K | $0.140 / $0.570 | |
| Healthy | 100%◆ (best) | 99.18% | 97.50% | 97.50% | 503ms◆ (best) | 5.75s | 32 tps | 41K | $0.140 / $0.400 | |
| Healthy | 99.98% | 99.87% | 99.86%◆ (best) | 99.77%◆ (best) | 519ms | 3.72s◆ (best) | 21 tps | 41K | $0.080 / $0.280 | |
Groqgroq | No data | — | — | — | — | — | — | — | 131K | $0.290 / $0.590 |
| No data | — | — | — | — | — | — | — | 41K | $0.100 / $0.300 | |
| No data | — | — | 95.56% | 95.56% | — | — | — | 41K | $0.070 / $0.200◆ (best) |
- SiliconFlow100%◆ (best)Healthy24h 99.27%
- 7d
- 98.69%
- 30d
- 97.46%
- p50
- 1.27s
- p99
- 14.3s
- Throughput
- 40 tps◆ (best)
- Context
- 131K
- $/Mtok
- $0.140 / $0.570
siliconflow/fp8fp8 - Tenstorrent100%◆ (best)Healthy24h 99.18%
- 7d
- 97.50%
- 30d
- 97.50%
- p50
- 503ms◆ (best)
- p99
- 5.75s
- Throughput
- 32 tps
- Context
- 41K
- $/Mtok
- $0.140 / $0.400
tenstorrent/fp8fp8 - DeepInfra99.98%Healthy24h 99.87%
- 7d
- 99.86%◆ (best)
- 30d
- 99.77%◆ (best)
- p50
- 519ms
- p99
- 3.72s◆ (best)
- Throughput
- 21 tps
- Context
- 41K
- $/Mtok
- $0.080 / $0.280
deepinfra/fp8fp8 - Groq—No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 131K
- $/Mtok
- $0.290 / $0.590
groq - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 41K
- $/Mtok
- $0.100 / $0.300
nebius/basefp8 - No data24h —
- 7d
- 95.56%
- 30d
- 95.56%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 41K
- $/Mtok
- $0.070 / $0.200◆ (best)
tenstorrent/bf16bf16
Measured 4m 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.