Kimi K2.6: 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Cloudflarecloudflare | Healthy | 100%◆ (best) | 99.98% | 87.22% | 87.22% | 1.12s | 8.10s | 35 tps | 262K | $0.950 / $4.00 |
| Healthy | 100%◆ (best) | 98.03% | 97.55% | 97.55% | 802ms | 5.23s | 67 tps | 262K | $0.587 / $2.47 | |
DigitalOceandigitalocean | Healthy | 100%◆ (best) | 99.77% | 99.14% | 99.14% | 827ms | 8.18s | 37 tps | 262K | $0.950 / $4.00 |
| Healthy | 100%◆ (best) | 99.68% | 99.44% | 99.44% | 687ms | 9.20s | 63 tps | 262K | $0.600 / $3.41 | |
| Healthy | 100%◆ (best) | 99.97% | 99.91%◆ (best) | 99.91%◆ (best) | 2.25s | 25.0s | 30 tps | 262K | $0.950 / $4.00 | |
Novitanovita | Healthy | 100%◆ (best) | 99.75% | 99.68% | 99.68% | 1.42s | 19.1s | 23 tps | 262K | $0.800 / $3.40 |
| Healthy | 100%◆ (best) | 99.59% | 99.10% | 99.10% | 893ms | 9.17s | 80 tps | 262K | $0.750 / $3.50 | |
| Healthy | 100%◆ (best) | 99.82% | 99.81% | 99.81% | 1.84s | 14.3s | 43 tps | 262K | $0.770 / $3.40 | |
| Healthy | 100%◆ (best) | 97.93% | 98.23% | 98.23% | 5.10s | 25.2s | 49 tps | 256K | $0.598 / $2.52 | |
| Healthy | 100%◆ (best) | 96.10% | 96.46% | 96.46% | 1.48s | 7.95s | 16 tps | 256K | $0.750 / $3.50 | |
| Healthy | 99.83% | 99.85% | 99.82% | 99.82% | 1.42s | 8.77s | 35 tps | 262K | $0.580 / $2.44◆ (best) | |
| Healthy | 99.74% | 99.36% | 99.42% | 99.42% | 854ms | 13.5s | 111 tps◆ (best) | 262K | $0.650 / $3.41 | |
| Healthy | 99.08% | 99.73% | 99.15% | 99.15% | 1.20s | 8.20s | 13 tps | 262K | $0.750 / $3.50 | |
| Healthy | 96.67% | 98.21% | 98.36% | 98.36% | 3.48s | 36.5s | 34 tps | 262K | $0.580 / $3.40 | |
| No data | — | 95.92% | 96.19% | 96.19% | 2.43s | 13.1s | 31 tps | 262K | $0.950 / $4.00 | |
| No data | — | — | — | — | 736ms | 1.31s◆ (best) | 24 tps | 262K | $0.950 / $4.00 | |
| Degraded | — | 91.15% | 88.17% | 88.17% | 1.10s | 114s | 4.0 tps | 262K | $0.700 / $3.50 | |
Fireworksfireworks | No data | — | 82.40% | 96.85% | 96.85% | 479ms◆ (best) | 5.76s | 76 tps | 262K | $0.950 / $4.00 |
Phalaphala | No data | — | 98.63% | 98.28% | 98.28% | 2.58s | 16.4s | 20 tps | 262K | $1.09 / $4.60 |
| No data | — | — | 99.88% | 99.88% | — | — | — | 262K | $1.00 / $4.00 | |
Togethertogether | No data | — | — | 97.92% | 97.92% | — | — | — | 262K | $1.20 / $4.50 |
- Cloudflare100%◆ (best)Healthy24h 99.98%
- 7d
- 87.22%
- 30d
- 87.22%
- p50
- 1.12s
- p99
- 8.10s
- Throughput
- 35 tps
- Context
- 262K
- $/Mtok
- $0.950 / $4.00
cloudflare - Decart100%◆ (best)Healthy24h 98.03%
- 7d
- 97.55%
- 30d
- 97.55%
- p50
- 802ms
- p99
- 5.23s
- Throughput
- 67 tps
- Context
- 262K
- $/Mtok
- $0.587 / $2.47
decart/fp4fp4 - DigitalOcean100%◆ (best)Healthy24h 99.77%
- 7d
- 99.14%
- 30d
- 99.14%
- p50
- 827ms
- p99
- 8.18s
- Throughput
- 37 tps
- Context
- 262K
- $/Mtok
- $0.950 / $4.00
digitalocean - Inceptron100%◆ (best)Healthy24h 99.68%
- 7d
- 99.44%
- 30d
- 99.44%
- p50
- 687ms
- p99
- 9.20s
- Throughput
- 63 tps
- Context
- 262K
- $/Mtok
- $0.600 / $3.41
inceptron/int4int4 - Moonshot AI100%◆ (best)Healthy24h 99.97%
- 7d
- 99.91%◆ (best)
- 30d
- 99.91%◆ (best)
- p50
- 2.25s
- p99
- 25.0s
- Throughput
- 30 tps
- Context
- 262K
- $/Mtok
- $0.950 / $4.00
moonshotai/int4int4 - Novita100%◆ (best)Healthy24h 99.75%
- 7d
- 99.68%
- 30d
- 99.68%
- p50
- 1.42s
- p99
- 19.1s
- Throughput
- 23 tps
- Context
- 262K
- $/Mtok
- $0.800 / $3.40
novita - Parasail100%◆ (best)Healthy24h 99.59%
- 7d
- 99.10%
- 30d
- 99.10%
- p50
- 893ms
- p99
- 9.17s
- Throughput
- 80 tps
- Context
- 262K
- $/Mtok
- $0.750 / $3.50
parasail/int4int4 - SiliconFlow100%◆ (best)Healthy24h 99.82%
- 7d
- 99.81%
- 30d
- 99.81%
- p50
- 1.84s
- p99
- 14.3s
- Throughput
- 43 tps
- Context
- 262K
- $/Mtok
- $0.770 / $3.40
siliconflow/fp8fp8 - StreamLake100%◆ (best)Healthy24h 97.93%
- 7d
- 98.23%
- 30d
- 98.23%
- p50
- 5.10s
- p99
- 25.2s
- Throughput
- 49 tps
- Context
- 256K
- $/Mtok
- $0.598 / $2.52
streamlake/fp8fp8 - Venice100%◆ (best)Healthy24h 96.10%
- 7d
- 96.46%
- 30d
- 96.46%
- p50
- 1.48s
- p99
- 7.95s
- Throughput
- 16 tps
- Context
- 256K
- $/Mtok
- $0.750 / $3.50
venice/int4int4 - Baidu99.83%Healthy24h 99.85%
- 7d
- 99.82%
- 30d
- 99.82%
- p50
- 1.42s
- p99
- 8.77s
- Throughput
- 35 tps
- Context
- 262K
- $/Mtok
- $0.580 / $2.44◆ (best)
baidu/fp4fp4 - CoreWeave99.74%Healthy24h 99.36%
- 7d
- 99.42%
- 30d
- 99.42%
- p50
- 854ms
- p99
- 13.5s
- Throughput
- 111 tps◆ (best)
- Context
- 262K
- $/Mtok
- $0.650 / $3.41
coreweave/fp4fp4 - DeepInfra99.08%Healthy24h 99.73%
- 7d
- 99.15%
- 30d
- 99.15%
- p50
- 1.20s
- p99
- 8.20s
- Throughput
- 13 tps
- Context
- 262K
- $/Mtok
- $0.750 / $3.50
deepinfra/fp4fp4 - Chutes96.67%Healthy24h 98.21%
- 7d
- 98.36%
- 30d
- 98.36%
- p50
- 3.48s
- p99
- 36.5s
- Throughput
- 34 tps
- Context
- 262K
- $/Mtok
- $0.580 / $3.40
chutes/int4int4 - No data24h 95.92%
- 7d
- 96.19%
- 30d
- 96.19%
- p50
- 2.43s
- p99
- 13.1s
- Throughput
- 31 tps
- Context
- 262K
- $/Mtok
- $0.950 / $4.00
atlas-cloud/int4int4 - No data24h —
- 7d
- —
- 30d
- —
- p50
- 736ms
- p99
- 1.31s◆ (best)
- Throughput
- 24 tps
- Context
- 262K
- $/Mtok
- $0.950 / $4.00
baseten/fp4fp4 - Degraded24h 91.15%
- 7d
- 88.17%
- 30d
- 88.17%
- p50
- 1.10s
- p99
- 114s
- Throughput
- 4.0 tps
- Context
- 262K
- $/Mtok
- $0.700 / $3.50
crusoe/bf16bf16 - No data24h 82.40%
- 7d
- 96.85%
- 30d
- 96.85%
- p50
- 479ms◆ (best)
- p99
- 5.76s
- Throughput
- 76 tps
- Context
- 262K
- $/Mtok
- $0.950 / $4.00
fireworks - No data24h 98.63%
- 7d
- 98.28%
- 30d
- 98.28%
- p50
- 2.58s
- p99
- 16.4s
- Throughput
- 20 tps
- Context
- 262K
- $/Mtok
- $1.09 / $4.60
phala - No data24h —
- 7d
- 99.88%
- 30d
- 99.88%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $1.00 / $4.00
sail-research/int4int4 - No data24h —
- 7d
- 97.92%
- 30d
- 97.92%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $1.20 / $4.50
together
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.