DeepSeek V4.1 Flash: 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Alibabaalibaba | Healthy | 100%◆ (best) | 100% | 99.71% | 99.05% | 566ms | 1.43s | 84 tps | 1.0M | $0.165 / $0.660 |
| Healthy | 100%◆ (best) | 100% | 99.74% | 99.74% | 723ms | 1.70s | 63 tps | 1.0M | $0.135 / $0.540 | |
| Healthy | 100%◆ (best) | 100% | 99.68% | 99.49% | 790ms | 3.36s | 91 tps | 1.0M | $0.300 / $1.20 | |
BaseTenbaseten | Healthy | 100%◆ (best) | 99.93% | 99.82% | 99.86% | 193ms◆ (best) | 1.97s | 176 tps | 1.0M | $0.300 / $1.20 |
BaseTenbaseten/us | Healthy | 100%◆ (best) | 99.79% | 99.72% | 99.83% | 289ms | 7.06s | 223 tps | 1.0M | $0.300 / $1.20 |
Coherecohere | Healthy | 100%◆ (best) | 100% | 99.78% | 99.77% | 260ms | 3.10s | 185 tps | 1.0M | $0.300 / $1.20 |
| Healthy | 100%◆ (best) | 99.93% | 98.92% | 98.61% | 631ms | 7.57s | 180 tps | 1.0M | $0.130 / $0.650 | |
| Healthy | 100%◆ (best) | 100% | 99.93% | 99.73% | 441ms | 2.35s | 144 tps | 1.0M | $0.290 / $1.20 | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.99% | 515ms | 3.02s | 111 tps | 1.0M | $0.140 / $0.420 | |
| Healthy | 100%◆ (best) | 100% | 98.87% | 95.85% | 1.89s | 5.36s | 118 tps | 1.0M | $0.112 / $0.336 | |
| Healthy | 100%◆ (best) | 100% | 71.20% | 71.30% | 494ms | 2.41s | 125 tps | 1.0M | $0.210 / $0.630 | |
DigitalOceandigitalocean | Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.88% | 521ms | 3.80s | 189 tps | 1.0M | $0.225 / $0.900 |
Fireworksfireworks/spot | Healthy | 100%◆ (best) | 100% | 99.85% | 99.91% | 637ms | 3.59s | 165 tps | 1.0M | $0.210 / $0.840 |
Fireworksfireworks/us | Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.36% | 248ms | 1.57s | 238 tps◆ (best) | 1.0M | $0.450 / $1.80 |
| Healthy | 100%◆ (best) | 98.78% | 96.46% | 98.87% | 1.33s | 5.17s | 47 tps | 1.0M | $0.180 / $0.720 | |
Googlegoogle-vertex/global | Degraded | 100%◆ (best) | 100% | 99.36% | 97.87% | 703ms | 15.2s | 105 tps | 1.0M | $0.300 / $1.20 |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.97% | 253ms | 1.69s | 118 tps | 1.0M | $0.103 / $0.600 | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.97% | 249ms | 1.12s | 80 tps | 1.0M | $0.103 / $0.600 | |
| Healthy | 100%◆ (best) | 99.71% | 99.75% | 99.48% | 247ms | 2.99s | 143 tps | 1.0M | $0.260 / $1.17 | |
Modalmodal | Healthy | 100%◆ (best) | 100% | 99.95% | 99.96% | 269ms | 1.10s◆ (best) | 158 tps | 1.0M | $0.300 / $1.20 |
| Healthy | 100%◆ (best) | 100% | 99.95% | 99.11% | 491ms | 2.68s | 93 tps | 1.0M | $0.050 / $0.192 | |
| Healthy | 100%◆ (best) | 97.96% | 99.02% | 98.95% | 696ms | 2.69s | 135 tps | 1.0M | $0.300 / $1.20 | |
| Healthy | 100%◆ (best) | 100% | 99.94% | 99.24% | 861ms | 2.53s | 109 tps | 1.0M | $0.195 / $0.780 | |
| Healthy | 100%◆ (best) | 99.72% | 99.65% | 98.35% | 1.01s | 14.0s | 97 tps | 1.0M | $0.035 / $0.500◆ (best) | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 98.84% | 239ms | 1.98s | 178 tps | 1.0M | $0.300 / $1.20 | |
PrimeIntellectprimeintellect | Healthy | 100%◆ (best) | 100% | 99.99% | 97.50% | 444ms | 2.32s | 184 tps | 1.0M | $0.300 / $1.20 |
Rekareka | Healthy | 100%◆ (best) | 100% | 99.24% | 97.37% | 345ms | 2.29s | 108 tps | 1.0M | $0.105 / $0.840 |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 100%◆ (best) | 572ms | 3.84s | 137 tps | 1.0M | $0.035 / $0.500◆ (best) | |
| Healthy | 100%◆ (best) | 99.93% | 99.95% | 99.82% | 1.57s | 4.20s | 112 tps | 1.0M | $0.300 / $1.20 | |
| Healthy | 100%◆ (best) | 100% | 99.98% | 99.86% | 1.55s | 6.99s | 75 tps | 1.0M | $0.147 / $0.588 | |
Togethertogether | Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.59% | 204ms | 3.17s | 229 tps | 1.0M | $0.300 / $1.20 |
| Healthy | 100%◆ (best) | 99.79% | 99.81% | 99.71% | 233ms | 5.84s | 146 tps | 1.0M | $0.300 / $1.20 | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 99.98% | 255ms | 1.56s | 216 tps | 1.0M | $0.190 / $0.650 | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 100%◆ (best) | 279ms | 2.06s | 209 tps | 1.0M | $0.045 / $0.500 | |
BaseTenbaseten/fast | No data | — | — | 42.00% | 83.80% | — | — | — | 1.0M | $0.600 / $2.40 |
| No data | — | — | — | 99.84% | — | — | — | 1.0M | $0.300 / $1.20 | |
| No data | — | — | — | 99.94% | — | — | — | 1.0M | $0.300 / $1.20 | |
| No data | — | — | — | 100%◆ (best) | — | — | — | 1.0M | $0.300 / $1.20 | |
Cloudflarecloudflare | No data | — | — | — | 99.63% | — | — | — | 1.0M | $0.100 / $0.900 |
| No data | — | — | 85.02% | 96.06% | — | — | — | 1.0M | $0.090 / $0.180 | |
| No data | — | — | — | — | — | — | — | 1.0M | $0.300 / $1.20 | |
DeepSeekdeepseek | No data | — | — | — | 99.29% | — | — | — | 1.0M | $0.150 / $0.600 |
DekaLLMdekallm | No data | — | — | — | 98.25% | — | — | — | 1.0M | $0.120 / $1.20 |
Fireworksfireworks | No data | — | 97.24% | 98.30% | 94.73% | 662ms | 4.71s | 136 tps | 1.0M | $0.300 / $1.20 |
Fireworksfireworks/flex | No data | — | — | — | 100%◆ (best) | — | — | — | 1.0M | $0.210 / $0.840 |
InferenceNetinference-net | No data | — | — | — | 99.52% | — | — | — | 1.0M | $0.070 / $0.300 |
| No data | — | — | — | — | — | — | — | 1.0M | $0.110 / $0.410 | |
Ionstreamionstream | No data | — | — | — | 99.15% | — | — | — | 1.0M | $0.220 / $0.600 |
| No data | — | — | — | — | — | — | — | 1.0M | $0.225 / $0.900 | |
Makoramakora | No data | — | — | — | — | — | — | — | 1.0M | $0.300 / $1.20 |
| No data | — | 98.08% | 97.88% | 94.86% | 1.02s | 3.68s | 99 tps | 1.0M | $0.200 / $1.00 | |
| No data | — | — | 100%◆ (best) | 96.25% | — | — | — | 1.0M | $0.170 / $0.670 | |
Morphmorph | No data | — | — | — | — | — | — | — | 1.0M | $0.079 / $0.317 |
| No data | — | — | — | 99.94% | — | — | — | 1.0M | $0.210 / $0.840 | |
Phalaphala | No data | — | 98.90% | 96.76% | 98.51% | 1.23s | 4.62s | 93 tps | 1.0M | $0.210 / $0.840 |
| No data | — | — | — | — | — | — | — | 262K | $0.290 / $1.16 | |
Relacerelace | No data | — | — | 99.99% | 99.79% | — | — | — | 1.0M | $0.070 / $0.500 |
| No data | — | — | 65.00% | 65.00% | — | — | — | 1.0M | $0.070 / $0.500 | |
| No data | — | — | 100%◆ (best) | 99.12% | — | — | — | 1.0M | $0.090 / $0.250 | |
| No data | — | — | 100%◆ (best) | 100%◆ (best) | — | — | — | 1.0M | $0.090 / $0.250 | |
StreamLakestreamlake | No data | — | — | — | — | — | — | — | 1.0M | $0.300 / $1.20 |
Waferwafer | No data | — | — | — | 99.83% | — | — | — | 1.0M | $0.100 / $0.650 |
- Alibaba100%◆ (best)Healthy24h 100%
- 7d
- 99.71%
- 30d
- 99.05%
- p50
- 566ms
- p99
- 1.43s
- Throughput
- 84 tps
- Context
- 1.0M
- $/Mtok
- $0.165 / $0.660
alibaba - AtlasCloud100%◆ (best)Healthy24h 100%
- 7d
- 99.74%
- 30d
- 99.74%
- p50
- 723ms
- p99
- 1.70s
- Throughput
- 63 tps
- Context
- 1.0M
- $/Mtok
- $0.135 / $0.540
atlas-cloud/fp8fp8 - Baidu100%◆ (best)Healthy24h 100%
- 7d
- 99.68%
- 30d
- 99.49%
- p50
- 790ms
- p99
- 3.36s
- Throughput
- 91 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
baidu/fp8fp8 - BaseTen100%◆ (best)Healthy24h 99.93%
- 7d
- 99.82%
- 30d
- 99.86%
- p50
- 193ms◆ (best)
- p99
- 1.97s
- Throughput
- 176 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
baseten - BaseTen100%◆ (best)Healthy24h 99.79%
- 7d
- 99.72%
- 30d
- 99.83%
- p50
- 289ms
- p99
- 7.06s
- Throughput
- 223 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
baseten/us - Cohere100%◆ (best)Healthy24h 100%
- 7d
- 99.78%
- 30d
- 99.77%
- p50
- 260ms
- p99
- 3.10s
- Throughput
- 185 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
cohere - CoreWeave100%◆ (best)Healthy24h 99.93%
- 7d
- 98.92%
- 30d
- 98.61%
- p50
- 631ms
- p99
- 7.57s
- Throughput
- 180 tps
- Context
- 1.0M
- $/Mtok
- $0.130 / $0.650
coreweave/fp8fp8 - Crusoe100%◆ (best)Healthy24h 100%
- 7d
- 99.93%
- 30d
- 99.73%
- p50
- 441ms
- p99
- 2.35s
- Throughput
- 144 tps
- Context
- 1.0M
- $/Mtok
- $0.290 / $1.20
crusoe/fp8fp8 - DeepInfra100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.99%
- p50
- 515ms
- p99
- 3.02s
- Throughput
- 111 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.420
deepinfra/fp8fp8 - DeepInfra100%◆ (best)Healthy24h 100%
- 7d
- 98.87%
- 30d
- 95.85%
- p50
- 1.89s
- p99
- 5.36s
- Throughput
- 118 tps
- Context
- 1.0M
- $/Mtok
- $0.112 / $0.336
deepinfra/fp8/flexfp8 - DeepInfra100%◆ (best)Healthy24h 100%
- 7d
- 71.20%
- 30d
- 71.30%
- p50
- 494ms
- p99
- 2.41s
- Throughput
- 125 tps
- Context
- 1.0M
- $/Mtok
- $0.210 / $0.630
deepinfra/fp8/priorityfp8 - DigitalOcean100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.88%
- p50
- 521ms
- p99
- 3.80s
- Throughput
- 189 tps
- Context
- 1.0M
- $/Mtok
- $0.225 / $0.900
digitalocean - Fireworks100%◆ (best)Healthy24h 100%
- 7d
- 99.85%
- 30d
- 99.91%
- p50
- 637ms
- p99
- 3.59s
- Throughput
- 165 tps
- Context
- 1.0M
- $/Mtok
- $0.210 / $0.840
fireworks/spot - Fireworks100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.36%
- p50
- 248ms
- p99
- 1.57s
- Throughput
- 238 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $0.450 / $1.80
fireworks/us - GMICloud100%◆ (best)Healthy24h 98.78%
- 7d
- 96.46%
- 30d
- 98.87%
- p50
- 1.33s
- p99
- 5.17s
- Throughput
- 47 tps
- Context
- 1.0M
- $/Mtok
- $0.180 / $0.720
gmicloud/fp8fp8 - Google100%◆ (best)Degraded24h 100%
- 7d
- 99.36%
- 30d
- 97.87%
- p50
- 703ms
- p99
- 15.2s
- Throughput
- 105 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
google-vertex/global - InferenceNet100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.97%
- p50
- 253ms
- p99
- 1.69s
- Throughput
- 118 tps
- Context
- 1.0M
- $/Mtok
- $0.103 / $0.600
inference-net/fp8fp8 - InferenceNet100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.97%
- p50
- 249ms
- p99
- 1.12s
- Throughput
- 80 tps
- Context
- 1.0M
- $/Mtok
- $0.103 / $0.600
inference-net/usfp8 - Makora100%◆ (best)Healthy24h 99.71%
- 7d
- 99.75%
- 30d
- 99.48%
- p50
- 247ms
- p99
- 2.99s
- Throughput
- 143 tps
- Context
- 1.0M
- $/Mtok
- $0.260 / $1.17
makora/fp8fp8 - Modal100%◆ (best)Healthy24h 100%
- 7d
- 99.95%
- 30d
- 99.96%
- p50
- 269ms
- p99
- 1.10s◆ (best)
- Throughput
- 158 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
modal - Morph100%◆ (best)Healthy24h 100%
- 7d
- 99.95%
- 30d
- 99.11%
- p50
- 491ms
- p99
- 2.68s
- Throughput
- 93 tps
- Context
- 1.0M
- $/Mtok
- $0.050 / $0.192
morph/fp8fp8 - Nebius100%◆ (best)Healthy24h 97.96%
- 7d
- 99.02%
- 30d
- 98.95%
- p50
- 696ms
- p99
- 2.69s
- Throughput
- 135 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
nebius/fp8fp8 - Novita100%◆ (best)Healthy24h 100%
- 7d
- 99.94%
- 30d
- 99.24%
- p50
- 861ms
- p99
- 2.53s
- Throughput
- 109 tps
- Context
- 1.0M
- $/Mtok
- $0.195 / $0.780
novita/fp8fp8 - OpenInference100%◆ (best)Healthy24h 99.72%
- 7d
- 99.65%
- 30d
- 98.35%
- p50
- 1.01s
- p99
- 14.0s
- Throughput
- 97 tps
- Context
- 1.0M
- $/Mtok
- $0.035 / $0.500◆ (best)
open-inference/fp4fp4 - Parasail100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 98.84%
- p50
- 239ms
- p99
- 1.98s
- Throughput
- 178 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
parasail/fp8fp8 - PrimeIntellect100%◆ (best)Healthy24h 100%
- 7d
- 99.99%
- 30d
- 97.50%
- p50
- 444ms
- p99
- 2.32s
- Throughput
- 184 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
primeintellect - Reka100%◆ (best)Healthy24h 100%
- 7d
- 99.24%
- 30d
- 97.37%
- p50
- 345ms
- p99
- 2.29s
- Throughput
- 108 tps
- Context
- 1.0M
- $/Mtok
- $0.105 / $0.840
reka - Relace100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 100%◆ (best)
- p50
- 572ms
- p99
- 3.84s
- Throughput
- 137 tps
- Context
- 1.0M
- $/Mtok
- $0.035 / $0.500◆ (best)
relace/fp4fp4 - SiliconFlow100%◆ (best)Healthy24h 99.93%
- 7d
- 99.95%
- 30d
- 99.82%
- p50
- 1.57s
- p99
- 4.20s
- Throughput
- 112 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
siliconflow/fp8fp8 - StreamLake100%◆ (best)Healthy24h 100%
- 7d
- 99.98%
- 30d
- 99.86%
- p50
- 1.55s
- p99
- 6.99s
- Throughput
- 75 tps
- Context
- 1.0M
- $/Mtok
- $0.147 / $0.588
streamlake/fp8fp8 - Together100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.59%
- p50
- 204ms
- p99
- 3.17s
- Throughput
- 229 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
together - Venice100%◆ (best)Healthy24h 99.79%
- 7d
- 99.81%
- 30d
- 99.71%
- p50
- 233ms
- p99
- 5.84s
- Throughput
- 146 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
venice/fp8fp8 - Wafer100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.98%
- p50
- 255ms
- p99
- 1.56s
- Throughput
- 216 tps
- Context
- 1.0M
- $/Mtok
- $0.190 / $0.650
wafer/mxfp4mxfp4 - Wafer100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 100%◆ (best)
- p50
- 279ms
- p99
- 2.06s
- Throughput
- 209 tps
- Context
- 1.0M
- $/Mtok
- $0.045 / $0.500
wafer/usmxfp4 - No data24h —
- 7d
- 42.00%
- 30d
- 83.80%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.600 / $2.40
baseten/fast - No data24h —
- 7d
- —
- 30d
- 99.84%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
baseten/fp32fp32 - No data24h —
- 7d
- —
- 30d
- 99.94%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
baseten/fp8fp8 - No data24h —
- 7d
- —
- 30d
- 100%◆ (best)
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
baseten/fp8/usfp8 - No data24h —
- 7d
- —
- 30d
- 99.63%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.100 / $0.900
cloudflare - No data24h —
- 7d
- 85.02%
- 30d
- 96.06%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.090 / $0.180
decart/fp4fp4 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
deepinfra/fp4fp4 - No data24h —
- 7d
- —
- 30d
- 99.29%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.150 / $0.600
deepseek - No data24h —
- 7d
- —
- 30d
- 98.25%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.120 / $1.20
dekallm - No data24h 97.24%
- 7d
- 98.30%
- 30d
- 94.73%
- p50
- 662ms
- p99
- 4.71s
- Throughput
- 136 tps
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
fireworks - No data24h —
- 7d
- —
- 30d
- 100%◆ (best)
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.210 / $0.840
fireworks/flex - No data24h —
- 7d
- —
- 30d
- 99.52%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.070 / $0.300
inference-net - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.110 / $0.410
io-net/fp8fp8 - No data24h —
- 7d
- —
- 30d
- 99.15%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.220 / $0.600
ionstream - Krea—No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.225 / $0.900
krea/fp8fp8 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
makora - No data24h 98.08%
- 7d
- 97.88%
- 30d
- 94.86%
- p50
- 1.02s
- p99
- 3.68s
- Throughput
- 99 tps
- Context
- 1.0M
- $/Mtok
- $0.200 / $1.00
mancer/fp8fp8 - No data24h —
- 7d
- 100%◆ (best)
- 30d
- 96.25%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.170 / $0.670
modelrun/fp4fp4 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.079 / $0.317
morph - No data24h —
- 7d
- —
- 30d
- 99.94%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.210 / $0.840
nextbit/fp8fp8 - No data24h 98.90%
- 7d
- 96.76%
- 30d
- 98.51%
- p50
- 1.23s
- p99
- 4.62s
- Throughput
- 93 tps
- Context
- 1.0M
- $/Mtok
- $0.210 / $0.840
phala - Reka—No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.290 / $1.16
reka/fp4fp4 - No data24h —
- 7d
- 99.99%
- 30d
- 99.79%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.070 / $0.500
relace - No data24h —
- 7d
- 65.00%
- 30d
- 65.00%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.070 / $0.500
relace/mxfp4mxfp4 - No data24h —
- 7d
- 100%◆ (best)
- 30d
- 99.12%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.090 / $0.250
sail-research/fp4fp4 - No data24h —
- 7d
- 100%◆ (best)
- 30d
- 100%◆ (best)
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.090 / $0.250
sail-research/fp4/usfp4 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.300 / $1.20
streamlake - No data24h —
- 7d
- —
- 30d
- 99.83%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.100 / $0.650
wafer
Measured 48s 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.