DeepSeek V4 Flash 0731
deepseek/deepseek-v4-flash-0731
Compare with DeepSeek V4 Flash 0423, DeepSeek V4 Pro 0423, DeepSeek V4 Pro 0813
CompareDeepSeek V4 Flash 0731
| Provider | State | Uptime 5m | 24h | p50 latency | p99 | Throughput | Context | $/Mtok |
|---|---|---|---|---|---|---|---|---|
BaseTenbaseten/fp8fp8 | Healthy | 100% | 98.34% | 584ms | 8.20s | 48 tps | 1.0M | $0.130 / $0.260 |
DeepSeekdeepseek | Healthy | 100% | 67.73% | 801ms | 2.70s | 73 tps | 1.0M | $0.220 / $0.660 |
Fireworksfireworks | Healthy | 100% | 93.07% | 794ms | 19.2s | 92 tps | 1.0M | $0.140 / $0.280 |
Novitanovita/fp8fp8 | Healthy | 99.99% | 99.33% | 995ms | 4.65s | 99 tps | 1.0M | $0.140 / $0.280 |
Baidubaidu/fp8fp8 | Healthy | 99.98% | 99.26% | 947ms | 7.47s | 103 tps | 1.0M | $0.111 / $0.221 |
CoreWeavecoreweave/fp8fp8 | Healthy | 99.97% | 99.89% | 666ms | 8.26s | 107 tps | 262K | $0.130 / $0.280 |
Decartdecart/fp4fp4 | Healthy | 99.94% | 73.28% | 822ms | 20.6s | 42 tps | 262K | $0.076 / $0.153 |
DigitalOceandigitalocean | Healthy | 99.93% | 95.71% | 658ms | 5.62s | 21 tps | 1.0M | $0.080 / $0.252 |
Cloudflarecloudflare | Healthy | 99.89% | 99.95% | 1.09s | 8.01s | 49 tps | 1.3M | $0.440 / $1.32 |
GMICloudgmicloud/fp8fp8 | Healthy | 99.84% | 98.61% | 2.64s | 12.3s | 56 tps | 1.0M | $0.112 / $0.224 |
AkashMLakashml/fp8fp8 | Healthy | 99.78% | 98.06% | 810ms | 9.35s | 28 tps | 131K | $0.140 / $0.280 |
Sail Researchsail-research/fp4fp4 | Healthy | 99.69% | 99.10% | 980ms | 40.2s | 28 tps | 1.0M | $0.090 / $0.180 |
Morphmorph/bf16bf16 | Healthy | 99.69% | 97.79% | 1.68s | 27.6s | 10 tps | 1.0M | $0.139 / $0.278 |
Inceptroninceptron/fp4fp4 | Healthy | 99.51% | 98.90% | 846ms | 17.9s | 23 tps | 1.0M | $0.130 / $0.280 |
Relacerelace/fp4fp4 | Healthy | 99.41% | 99.31% | 531ms | 3.88s | 74 tps | 1.0M | $0.105 / $0.210 |
Ambientambient/fp4fp4 | Degraded | 99.40% | 92.39% | 1.34s | 24.7s | 26 tps | 1.0M | $0.140 / $0.280 |
AtlasCloudatlas-cloud/fp4fp4 | Healthy | 99.39% | 99.32% | 2.17s | 11.1s | 96 tps | 1.0M | $0.140 / $0.280 |
DeepInfradeepinfra/fp8fp8 | Healthy | 99.28% | 98.07% | 858ms | 21.2s | 25 tps | 1.0M | $0.080 / $0.180 |
SiliconFlowsiliconflow/fp8fp8 | Healthy | 99.26% | 98.04% | 1.68s | 25.3s | 64 tps | 1.0M | $0.140 / $0.280 |
Waferwafer/fast | Healthy | 98.87% | 96.65% | 1.29s | 27.9s | 103 tps | 1.0M | $0.280 / $0.560 |
Venicevenice | Healthy | 98.83% | 95.94% | 928ms | 13.8s | 49 tps | 1.0M | $0.175 / $0.350 |
Parasailparasail/fp8fp8 | Healthy | 98.77% | 98.50% | 787ms | 9.78s | 56 tps | 1.0M | $0.140 / $0.280 |
Togethertogether | Healthy | 98.03% | 96.57% | 786ms | 11.4s | 46 tps | 1.0M | $0.140 / $0.280 |
OpenInferenceopen-inference/fp4fp4 | Healthy | 98.01% | 95.65% | 4.96s | 35.4s | 8.0 tps | 262K | $0.080 / $0.180 |
Io Netio-net/fp8fp8 | Healthy | 97.35% | 98.67% | 1.68s | 21.3s | 21 tps | 262K | $0.149 / $0.320 |
StreamLakestreamlake/fp8fp8 | Degraded | 96.75% | 95.23% | 2.67s | 33.7s | 38 tps | 1.0M | $0.079 / $0.157 |
Phalaphala | Degraded | 94.68% | 96.59% | 2.26s | 25.8s | 35 tps | 1.0M | $0.200 / $0.400 |
Mancer 2mancer/fp8fp8 | Degraded | 92.88% | 95.22% | 1.06s | 15.3s | 28 tps | 1.0M | $0.145 / $0.450 |
- BaseTen100%Healthy24h 98.34%
- p50
- 584ms
- p99
- 8.20s
- Throughput
- 48 tps
- Context
- 1.0M
- $/Mtok
- $0.130 / $0.260
baseten/fp8fp8 - DeepSeek100%Healthy24h 67.73%
- p50
- 801ms
- p99
- 2.70s
- Throughput
- 73 tps
- Context
- 1.0M
- $/Mtok
- $0.220 / $0.660
deepseek - Fireworks100%Healthy24h 93.07%
- p50
- 794ms
- p99
- 19.2s
- Throughput
- 92 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
fireworks - Novita99.99%Healthy24h 99.33%
- p50
- 995ms
- p99
- 4.65s
- Throughput
- 99 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
novita/fp8fp8 - Baidu99.98%Healthy24h 99.26%
- p50
- 947ms
- p99
- 7.47s
- Throughput
- 103 tps
- Context
- 1.0M
- $/Mtok
- $0.111 / $0.221
baidu/fp8fp8 - CoreWeave99.97%Healthy24h 99.89%
- p50
- 666ms
- p99
- 8.26s
- Throughput
- 107 tps
- Context
- 262K
- $/Mtok
- $0.130 / $0.280
coreweave/fp8fp8 - Decart99.94%Healthy24h 73.28%
- p50
- 822ms
- p99
- 20.6s
- Throughput
- 42 tps
- Context
- 262K
- $/Mtok
- $0.076 / $0.153
decart/fp4fp4 - DigitalOcean99.93%Healthy24h 95.71%
- p50
- 658ms
- p99
- 5.62s
- Throughput
- 21 tps
- Context
- 1.0M
- $/Mtok
- $0.080 / $0.252
digitalocean - Cloudflare99.89%Healthy24h 99.95%
- p50
- 1.09s
- p99
- 8.01s
- Throughput
- 49 tps
- Context
- 1.3M
- $/Mtok
- $0.440 / $1.32
cloudflare - GMICloud99.84%Healthy24h 98.61%
- p50
- 2.64s
- p99
- 12.3s
- Throughput
- 56 tps
- Context
- 1.0M
- $/Mtok
- $0.112 / $0.224
gmicloud/fp8fp8 - AkashML99.78%Healthy24h 98.06%
- p50
- 810ms
- p99
- 9.35s
- Throughput
- 28 tps
- Context
- 131K
- $/Mtok
- $0.140 / $0.280
akashml/fp8fp8 - Sail Research99.69%Healthy24h 99.10%
- p50
- 980ms
- p99
- 40.2s
- Throughput
- 28 tps
- Context
- 1.0M
- $/Mtok
- $0.090 / $0.180
sail-research/fp4fp4 - Morph99.69%Healthy24h 97.79%
- p50
- 1.68s
- p99
- 27.6s
- Throughput
- 10 tps
- Context
- 1.0M
- $/Mtok
- $0.139 / $0.278
morph/bf16bf16 - Inceptron99.51%Healthy24h 98.90%
- p50
- 846ms
- p99
- 17.9s
- Throughput
- 23 tps
- Context
- 1.0M
- $/Mtok
- $0.130 / $0.280
inceptron/fp4fp4 - Relace99.41%Healthy24h 99.31%
- p50
- 531ms
- p99
- 3.88s
- Throughput
- 74 tps
- Context
- 1.0M
- $/Mtok
- $0.105 / $0.210
relace/fp4fp4 - Ambient99.40%Degraded24h 92.39%
- p50
- 1.34s
- p99
- 24.7s
- Throughput
- 26 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
ambient/fp4fp4 - AtlasCloud99.39%Healthy24h 99.32%
- p50
- 2.17s
- p99
- 11.1s
- Throughput
- 96 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
atlas-cloud/fp4fp4 - DeepInfra99.28%Healthy24h 98.07%
- p50
- 858ms
- p99
- 21.2s
- Throughput
- 25 tps
- Context
- 1.0M
- $/Mtok
- $0.080 / $0.180
deepinfra/fp8fp8 - SiliconFlow99.26%Healthy24h 98.04%
- p50
- 1.68s
- p99
- 25.3s
- Throughput
- 64 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
siliconflow/fp8fp8 - Wafer98.87%Healthy24h 96.65%
- p50
- 1.29s
- p99
- 27.9s
- Throughput
- 103 tps
- Context
- 1.0M
- $/Mtok
- $0.280 / $0.560
wafer/fast - Venice98.83%Healthy24h 95.94%
- p50
- 928ms
- p99
- 13.8s
- Throughput
- 49 tps
- Context
- 1.0M
- $/Mtok
- $0.175 / $0.350
venice - Parasail98.77%Healthy24h 98.50%
- p50
- 787ms
- p99
- 9.78s
- Throughput
- 56 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
parasail/fp8fp8 - Together98.03%Healthy24h 96.57%
- p50
- 786ms
- p99
- 11.4s
- Throughput
- 46 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
together - OpenInference98.01%Healthy24h 95.65%
- p50
- 4.96s
- p99
- 35.4s
- Throughput
- 8.0 tps
- Context
- 262K
- $/Mtok
- $0.080 / $0.180
open-inference/fp4fp4 - Io Net97.35%Healthy24h 98.67%
- p50
- 1.68s
- p99
- 21.3s
- Throughput
- 21 tps
- Context
- 262K
- $/Mtok
- $0.149 / $0.320
io-net/fp8fp8 - StreamLake96.75%Degraded24h 95.23%
- p50
- 2.67s
- p99
- 33.7s
- Throughput
- 38 tps
- Context
- 1.0M
- $/Mtok
- $0.079 / $0.157
streamlake/fp8fp8 - Phala94.68%Degraded24h 96.59%
- p50
- 2.26s
- p99
- 25.8s
- Throughput
- 35 tps
- Context
- 1.0M
- $/Mtok
- $0.200 / $0.400
phala - Mancer 292.88%Degraded24h 95.22%
- p50
- 1.06s
- p99
- 15.3s
- Throughput
- 28 tps
- Context
- 1.0M
- $/Mtok
- $0.145 / $0.450
mancer/fp8fp8
Measured 5m ago. Latency and throughput are OpenRouter's rolling 30-minute windows.
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 and p99 are OpenRouter's rolling 30-minute windows.
Throughput
Median tokens per second.