Kimi K3
moonshotai/kimi-k3
Compare with Kimi K2.7 Code, Kimi K2.6, Kimi K2.5
CompareKimi K3
| State | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Healthy | 100% | 98.46% | 98.25% | 98.25% | 483ms | 14.5s | 87 tps | 1.0M | $3.00 / $15.00 | |
| Healthy | 100% | 99.28% | 97.21% | 97.21% | 1.84s | 77.0s | 28 tps | 1.0M | $3.00 / $15.00 | |
Fireworksfireworks | Healthy | 100% | 98.44% | 98.93% | 98.93% | 3.11s | 24.3s | 41 tps | 1.0M | $3.00 / $15.00 |
| Healthy | 99.80% | 99.53% | 99.65% | 99.65% | 1.65s | 16.5s | 77 tps | 1.0M | $3.00 / $15.00 | |
DigitalOceandigitalocean | Healthy | 99.39% | 99.45% | 99.31% | 99.31% | 3.17s | 30.9s | 24 tps | 1.0M | $2.85 / $14.25 |
Togethertogether | Healthy | 99.27% | 98.88% | 98.72% | 98.72% | 2.51s | 34.2s | 37 tps | 1.0M | $3.00 / $15.00 |
| Healthy | 98.78% | 95.95% | 88.21% | 88.21% | 3.25s | 50.8s | 26 tps | 1.0M | $2.80 / $14.00 | |
| Healthy | 98.71% | 97.39% | 98.48% | 98.48% | 3.19s | 112s | 34 tps | 975K | $2.60 / $13.00 | |
| Degraded | 98.65% | 98.31% | 99.15% | 99.15% | 7.76s | 48.0s | 19 tps | 1.0M | $3.00 / $15.00 | |
Fireworksfireworks/us | Healthy | 97.69% | 92.99% | 95.44% | 95.44% | 1.13s | 23.4s | 57 tps | 1.0M | $3.30 / $16.50 |
| Down | — | 90.47% | 88.82% | 88.82% | 1.09s | 214s | 19 tps | 1.0M | $2.85 / $14.25 | |
| No data | — | 84.52% | 82.82% | 82.82% | 2.34s | 21.7s | 27 tps | 1.0M | $6.00 / $22.50 | |
Phalaphala | No data | — | 95.67% | 94.92% | 94.92% | 5.73s | 35.9s | 22 tps | 1.0M | $3.00 / $15.00 |
Waferwafer | No data | — | 96.81% | 96.96% | 96.96% | — | — | — | 1.0M | $3.00 / $15.00 |
- BaseTen100%Healthy24h 98.46%
- 7d
- 98.25%
- 30d
- 98.25%
- p50
- 483ms
- p99
- 14.5s
- Throughput
- 87 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
baseten/fp8fp8 - Chutes100%Healthy24h 99.28%
- 7d
- 97.21%
- 30d
- 97.21%
- p50
- 1.84s
- p99
- 77.0s
- Throughput
- 28 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
chutes/mxfp4mxfp4 - Fireworks100%Healthy24h 98.44%
- 7d
- 98.93%
- 30d
- 98.93%
- p50
- 3.11s
- p99
- 24.3s
- Throughput
- 41 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
fireworks - Modal99.80%Healthy24h 99.53%
- 7d
- 99.65%
- 30d
- 99.65%
- p50
- 1.65s
- p99
- 16.5s
- Throughput
- 77 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
modal/mxfp4mxfp4 - DigitalOcean99.39%Healthy24h 99.45%
- 7d
- 99.31%
- 30d
- 99.31%
- p50
- 3.17s
- p99
- 30.9s
- Throughput
- 24 tps
- Context
- 1.0M
- $/Mtok
- $2.85 / $14.25
digitalocean - Together99.27%Healthy24h 98.88%
- 7d
- 98.72%
- 30d
- 98.72%
- p50
- 2.51s
- p99
- 34.2s
- Throughput
- 37 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
together - Morph98.78%Healthy24h 95.95%
- 7d
- 88.21%
- 30d
- 88.21%
- p50
- 3.25s
- p99
- 50.8s
- Throughput
- 26 tps
- Context
- 1.0M
- $/Mtok
- $2.80 / $14.00
morph/fp4fp4 - Sail Research98.71%Healthy24h 97.39%
- 7d
- 98.48%
- 30d
- 98.48%
- p50
- 3.19s
- p99
- 112s
- Throughput
- 34 tps
- Context
- 975K
- $/Mtok
- $2.60 / $13.00
sail-research/fp4fp4 - Moonshot AI98.65%Degraded24h 98.31%
- 7d
- 99.15%
- 30d
- 99.15%
- p50
- 7.76s
- p99
- 48.0s
- Throughput
- 19 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
moonshotai/mxfp4mxfp4 - Fireworks97.69%Healthy24h 92.99%
- 7d
- 95.44%
- 30d
- 95.44%
- p50
- 1.13s
- p99
- 23.4s
- Throughput
- 57 tps
- Context
- 1.0M
- $/Mtok
- $3.30 / $16.50
fireworks/us - Down24h 90.47%
- 7d
- 88.82%
- 30d
- 88.82%
- p50
- 1.09s
- p99
- 214s
- Throughput
- 19 tps
- Context
- 1.0M
- $/Mtok
- $2.85 / $14.25
deepinfra/bf16bf16 - No data24h 84.52%
- 7d
- 82.82%
- 30d
- 82.82%
- p50
- 2.34s
- p99
- 21.7s
- Throughput
- 27 tps
- Context
- 1.0M
- $/Mtok
- $6.00 / $22.50
morph/fastfp4 - No data24h 95.67%
- 7d
- 94.92%
- 30d
- 94.92%
- p50
- 5.73s
- p99
- 35.9s
- Throughput
- 22 tps
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
phala - No data24h 96.81%
- 7d
- 96.96%
- 30d
- 96.96%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $3.00 / $15.00
wafer
Measured 55s ago. Latency and throughput are OpenRouter's rolling 30-minute windows.
History
3 endpoints reporting no data are omitted from the charts.
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.