Kimi K2.5: 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) | 99.92% | 99.65% | 99.65% | 673ms◆ (best) | 17.1s | 51 tps | 262K | $0.450 / $2.25◆ (best) | |
DigitalOceandigitalocean | Healthy | 100%◆ (best) | 99.73% | 99.37% | 99.37% | 891ms | 4.63s | 12 tps | 262K | $0.500 / $2.70 |
| Healthy | 100%◆ (best) | 99.95% | 99.97% | 99.97% | 1.43s | 8.00s | 49 tps | 262K | $0.600 / $3.00 | |
| Healthy | 100%◆ (best) | 99.98% | 99.98%◆ (best) | 99.98%◆ (best) | 1.80s | 48.9s | 39 tps | 262K | $0.450 / $2.25◆ (best) | |
| Healthy | 100%◆ (best) | 99.32% | 99.61% | 99.61% | 1.26s | 12.1s | 55 tps | 256K | $0.540 / $2.70 | |
| Healthy | 99.77% | 99.35% | 99.37% | 99.37% | 1.33s | 11.4s | 39 tps | 262K | $0.490 / $2.50 | |
Venicevenice | Healthy | 99.76% | 97.62% | 98.71% | 98.71% | 1.66s | 7.64s | 107 tps◆ (best) | 256K | $0.532 / $3.32 |
Novitanovita | Healthy | 99.52% | 99.01% | 99.11% | 99.11% | 1.25s | 8.50s | 40 tps | 262K | $0.570 / $2.85 |
Amazon Bedrockamazon-bedrock/us-east-2 | No data | — | 99.03% | 98.72% | 98.72% | 998ms | 7.33s | 44 tps | 262K | $0.600 / $3.00 |
Phalaphala | No data | — | 80.76% | 94.12% | 94.12% | 1.19s | 1.87s◆ (best) | 53 tps | 262K | $0.600 / $3.00 |
- DeepInfra100%◆ (best)Healthy24h 99.92%
- 7d
- 99.65%
- 30d
- 99.65%
- p50
- 673ms◆ (best)
- p99
- 17.1s
- Throughput
- 51 tps
- Context
- 262K
- $/Mtok
- $0.450 / $2.25◆ (best)
deepinfra/fp4fp4 - DigitalOcean100%◆ (best)Healthy24h 99.73%
- 7d
- 99.37%
- 30d
- 99.37%
- p50
- 891ms
- p99
- 4.63s
- Throughput
- 12 tps
- Context
- 262K
- $/Mtok
- $0.500 / $2.70
digitalocean - Moonshot AI100%◆ (best)Healthy24h 99.95%
- 7d
- 99.97%
- 30d
- 99.97%
- p50
- 1.43s
- p99
- 8.00s
- Throughput
- 49 tps
- Context
- 262K
- $/Mtok
- $0.600 / $3.00
moonshotai/int4int4 - SiliconFlow100%◆ (best)Healthy24h 99.98%
- 7d
- 99.98%◆ (best)
- 30d
- 99.98%◆ (best)
- p50
- 1.80s
- p99
- 48.9s
- Throughput
- 39 tps
- Context
- 262K
- $/Mtok
- $0.450 / $2.25◆ (best)
siliconflow/int4int4 - StreamLake100%◆ (best)Healthy24h 99.32%
- 7d
- 99.61%
- 30d
- 99.61%
- p50
- 1.26s
- p99
- 12.1s
- Throughput
- 55 tps
- Context
- 256K
- $/Mtok
- $0.540 / $2.70
streamlake/fp8fp8 - AtlasCloud99.77%Healthy24h 99.35%
- 7d
- 99.37%
- 30d
- 99.37%
- p50
- 1.33s
- p99
- 11.4s
- Throughput
- 39 tps
- Context
- 262K
- $/Mtok
- $0.490 / $2.50
atlas-cloud/int4int4 - Venice99.76%Healthy24h 97.62%
- 7d
- 98.71%
- 30d
- 98.71%
- p50
- 1.66s
- p99
- 7.64s
- Throughput
- 107 tps◆ (best)
- Context
- 256K
- $/Mtok
- $0.532 / $3.32
venice - Novita99.52%Healthy24h 99.01%
- 7d
- 99.11%
- 30d
- 99.11%
- p50
- 1.25s
- p99
- 8.50s
- Throughput
- 40 tps
- Context
- 262K
- $/Mtok
- $0.570 / $2.85
novita - No data24h 99.03%
- 7d
- 98.72%
- 30d
- 98.72%
- p50
- 998ms
- p99
- 7.33s
- Throughput
- 44 tps
- Context
- 262K
- $/Mtok
- $0.600 / $3.00
amazon-bedrock/us-east-2 - No data24h 80.76%
- 7d
- 94.12%
- 30d
- 94.12%
- p50
- 1.19s
- p99
- 1.87s◆ (best)
- Throughput
- 53 tps
- Context
- 262K
- $/Mtok
- $0.600 / $3.00
phala
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.
Throughput
Median tokens per second.