DeepSeek V4 Pro 0813: 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) | 99.34% | 98.95% | 99.39% | 1.52s | 6.99s | 59 tps | 1.0M | $1.32 / $3.96 | |
| Healthy | 100%◆ (best) | 99.95% | 99.92% | 99.93% | 1.25s | 17.0s | 51 tps | 1.0M | $1.32 / $3.96 | |
| Healthy | 100%◆ (best) | 99.47% | 99.55% | 99.65% | 590ms◆ (best) | 4.04s | 113 tps | 1.0M | $1.31 / $3.96 | |
| Healthy | 100%◆ (best) | 99.09% | 99.47% | 99.25% | 1.42s | 26.9s | 45 tps | 1.0M | $1.30 / $2.60 | |
DeepSeekdeepseek | Healthy | 100%◆ (best) | 99.98% | 99.97% | 99.97%◆ (best) | 1.06s | 4.43s | 36 tps | 1.0M | $0.660 / $1.98 |
DigitalOceandigitalocean | Healthy | 100%◆ (best) | 99.97% | 99.89% | 99.70% | 674ms | 3.00s◆ (best) | 47 tps | 1.0M | $1.32 / $3.96 |
| Healthy | 100%◆ (best) | 99.87% | 99.95% | 99.77% | 3.85s | 32.4s | 31 tps | 1.0M | $1.06 / $3.17 | |
Ionstreamionstream | Healthy | 100%◆ (best) | 99.43% | 98.48% | 98.90% | 2.56s | 31.3s | 27 tps | 1.0M | $0.380 / $1.98 |
| Healthy | 100%◆ (best) | 99.16% | 98.81% | 87.62% | 1.31s | 31.2s | 100 tps | 979K | $1.32 / $3.96 | |
| Healthy | 100%◆ (best) | 99.93% | 99.95% | 99.47% | 3.11s | 7.21s | 48 tps | 1.0M | $1.06 / $3.17 | |
| Healthy | 100%◆ (best) | 99.87% | 99.96% | 99.92% | 1.14s | 4.43s | 97 tps | 1.0M | $0.990 / $2.97 | |
| Healthy | 100%◆ (best) | 94.20% | 97.46% | 99.15% | 1.12s | 22.1s | 40 tps | 1.0M | $1.32 / $3.96 | |
Phalaphala | Healthy | 100%◆ (best) | 98.22% | 99.15% | 99.46% | 4.01s | 13.3s | 41 tps | 1.0M | $0.957 / $2.88 |
| Healthy | 100%◆ (best) | 99.92% | 98.96% | 98.64% | 782ms | 4.35s | 120 tps◆ (best) | 1.0M | $0.300 / $4.20◆ (best) | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.18% | 1.51s | 12.7s | 64 tps | 1.0M | $1.32 / $3.96 | |
StreamLakestreamlake | Healthy | 100%◆ (best) | 99.46% | 99.62% | 99.71% | 2.27s | 14.9s | 43 tps | 1.0M | $0.660 / $1.98 |
Togethertogether | Healthy | 100%◆ (best) | 97.78% | 98.40% | 99.07% | 977ms | 27.6s | 65 tps | 1.0M | $1.32 / $3.96 |
Waferwafer | Healthy | 99.91% | 99.61% | 99.18% | 99.20% | 702ms | 7.83s | 113 tps | 1.0M | $0.400 / $5.00 |
Alibabaalibaba | Healthy | 97.60% | 96.89% | 98.15% | 99.31% | 1.37s | 3.95s | 72 tps | 1.0M | $1.12 / $3.37 |
| No data | — | — | — | 98.89% | — | — | — | 1.0M | $1.32 / $3.96 | |
Cloudflarecloudflare | No data | — | 99.25% | 99.39% | 98.26% | 1.18s | 6.01s | 38 tps | 1.0M | $1.32 / $3.96 |
Fireworksfireworks | No data | — | — | — | 96.19% | — | — | — | 1.0M | $1.32 / $3.96 |
| No data | — | — | 99.42% | 99.59% | — | — | — | 262K | $0.490 / $2.45 | |
| No data | — | — | — | 99.94% | — | — | — | 1.0M | $1.10 / $3.30 | |
Rekareka | No data | — | — | 99.51% | 99.51% | — | — | — | 1.0M | $1.12 / $3.37 |
| No data | — | — | 99.02% | 99.40% | — | — | — | 1.0M | $0.400 / $4.30 | |
| No data | — | — | 97.79% | 99.09% | — | — | — | 1.0M | $0.400 / $4.30 | |
Venicevenice | No data | — | 90.29% | 94.82% | 97.62% | 1.42s | 7.22s | 46 tps | 1.0M | $1.65 / $4.95 |
- AtlasCloud100%◆ (best)Healthy24h 99.34%
- 7d
- 98.95%
- 30d
- 99.39%
- p50
- 1.52s
- p99
- 6.99s
- Throughput
- 59 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
atlas-cloud/fp8fp8 - Baidu100%◆ (best)Healthy24h 99.95%
- 7d
- 99.92%
- 30d
- 99.93%
- p50
- 1.25s
- p99
- 17.0s
- Throughput
- 51 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
baidu/fp8fp8 - CoreWeave100%◆ (best)Healthy24h 99.47%
- 7d
- 99.55%
- 30d
- 99.65%
- p50
- 590ms◆ (best)
- p99
- 4.04s
- Throughput
- 113 tps
- Context
- 1.0M
- $/Mtok
- $1.31 / $3.96
coreweave/fp8fp8 - DeepInfra100%◆ (best)Healthy24h 99.09%
- 7d
- 99.47%
- 30d
- 99.25%
- p50
- 1.42s
- p99
- 26.9s
- Throughput
- 45 tps
- Context
- 1.0M
- $/Mtok
- $1.30 / $2.60
deepinfra/fp8fp8 - DeepSeek100%◆ (best)Healthy24h 99.98%
- 7d
- 99.97%
- 30d
- 99.97%◆ (best)
- p50
- 1.06s
- p99
- 4.43s
- Throughput
- 36 tps
- Context
- 1.0M
- $/Mtok
- $0.660 / $1.98
deepseek - DigitalOcean100%◆ (best)Healthy24h 99.97%
- 7d
- 99.89%
- 30d
- 99.70%
- p50
- 674ms
- p99
- 3.00s◆ (best)
- Throughput
- 47 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
digitalocean - GMICloud100%◆ (best)Healthy24h 99.87%
- 7d
- 99.95%
- 30d
- 99.77%
- p50
- 3.85s
- p99
- 32.4s
- Throughput
- 31 tps
- Context
- 1.0M
- $/Mtok
- $1.06 / $3.17
gmicloud/fp8fp8 - Ionstream100%◆ (best)Healthy24h 99.43%
- 7d
- 98.48%
- 30d
- 98.90%
- p50
- 2.56s
- p99
- 31.3s
- Throughput
- 27 tps
- Context
- 1.0M
- $/Mtok
- $0.380 / $1.98
ionstream - Nebius100%◆ (best)Healthy24h 99.16%
- 7d
- 98.81%
- 30d
- 87.62%
- p50
- 1.31s
- p99
- 31.2s
- Throughput
- 100 tps
- Context
- 979K
- $/Mtok
- $1.32 / $3.96
nebius/fp8fp8 - NextBit100%◆ (best)Healthy24h 99.93%
- 7d
- 99.95%
- 30d
- 99.47%
- p50
- 3.11s
- p99
- 7.21s
- Throughput
- 48 tps
- Context
- 1.0M
- $/Mtok
- $1.06 / $3.17
nextbit/fp8fp8 - Novita100%◆ (best)Healthy24h 99.87%
- 7d
- 99.96%
- 30d
- 99.92%
- p50
- 1.14s
- p99
- 4.43s
- Throughput
- 97 tps
- Context
- 1.0M
- $/Mtok
- $0.990 / $2.97
novita/fp8fp8 - Parasail100%◆ (best)Healthy24h 94.20%
- 7d
- 97.46%
- 30d
- 99.15%
- p50
- 1.12s
- p99
- 22.1s
- Throughput
- 40 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
parasail/fp8fp8 - Phala100%◆ (best)Healthy24h 98.22%
- 7d
- 99.15%
- 30d
- 99.46%
- p50
- 4.01s
- p99
- 13.3s
- Throughput
- 41 tps
- Context
- 1.0M
- $/Mtok
- $0.957 / $2.88
phala - Relace100%◆ (best)Healthy24h 99.92%
- 7d
- 98.96%
- 30d
- 98.64%
- p50
- 782ms
- p99
- 4.35s
- Throughput
- 120 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $0.300 / $4.20◆ (best)
relace/fp4fp4 - SiliconFlow100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.18%
- p50
- 1.51s
- p99
- 12.7s
- Throughput
- 64 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
siliconflow/fp8fp8 - StreamLake100%◆ (best)Healthy24h 99.46%
- 7d
- 99.62%
- 30d
- 99.71%
- p50
- 2.27s
- p99
- 14.9s
- Throughput
- 43 tps
- Context
- 1.0M
- $/Mtok
- $0.660 / $1.98
streamlake - Together100%◆ (best)Healthy24h 97.78%
- 7d
- 98.40%
- 30d
- 99.07%
- p50
- 977ms
- p99
- 27.6s
- Throughput
- 65 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
together - Wafer99.91%Healthy24h 99.61%
- 7d
- 99.18%
- 30d
- 99.20%
- p50
- 702ms
- p99
- 7.83s
- Throughput
- 113 tps
- Context
- 1.0M
- $/Mtok
- $0.400 / $5.00
wafer - Alibaba97.60%Healthy24h 96.89%
- 7d
- 98.15%
- 30d
- 99.31%
- p50
- 1.37s
- p99
- 3.95s
- Throughput
- 72 tps
- Context
- 1.0M
- $/Mtok
- $1.12 / $3.37
alibaba - No data24h —
- 7d
- —
- 30d
- 98.89%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
baseten/fp4fp4 - No data24h 99.25%
- 7d
- 99.39%
- 30d
- 98.26%
- p50
- 1.18s
- p99
- 6.01s
- Throughput
- 38 tps
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
cloudflare - No data24h —
- 7d
- —
- 30d
- 96.19%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $1.32 / $3.96
fireworks - No data24h —
- 7d
- 99.42%
- 30d
- 99.59%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.490 / $2.45
io-net/fp8fp8 - No data24h —
- 7d
- —
- 30d
- 99.94%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $1.10 / $3.30
ionstream/fp4fp4 - Reka—No data24h —
- 7d
- 99.51%
- 30d
- 99.51%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $1.12 / $3.37
reka - No data24h —
- 7d
- 99.02%
- 30d
- 99.40%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.400 / $4.30
sail-research/fp4fp4 - No data24h —
- 7d
- 97.79%
- 30d
- 99.09%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.400 / $4.30
sail-research/usfp4 - No data24h 90.29%
- 7d
- 94.82%
- 30d
- 97.62%
- p50
- 1.42s
- p99
- 7.22s
- Throughput
- 46 tps
- Context
- 1.0M
- $/Mtok
- $1.65 / $4.95
venice
Measured 1m 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.