MiMo-V2.5: 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 | 99.95%◆ (best) | 99.63% | 98.33% | 98.33% | 381ms◆ (best) | 9.59s | 70 tps◆ (best) | 1.0M | $0.140 / $0.280 | |
| Healthy | 99.85% | 99.67% | 99.04%◆ (best) | 99.04%◆ (best) | 2.70s | 28.6s | 36 tps | 1.0M | $0.140 / $0.280 | |
| Healthy | 99.84% | 97.85% | 98.59% | 98.59% | 2.22s | 14.9s | 39 tps | 1.0M | $0.168 / $0.336 | |
| Healthy | 98.70% | 98.56% | 95.55% | 95.55% | 533ms | 2.01s◆ (best) | 26 tps | 262K | $0.400 / $2.00 | |
| Degraded | 97.79% | 87.12% | 88.28% | 88.28% | 4.15s | 8.67s | 3.0 tps | 1.1M | $0.119 / $0.238◆ (best) |
- Parasail99.95%◆ (best)Healthy24h 99.63%
- 7d
- 98.33%
- 30d
- 98.33%
- p50
- 381ms◆ (best)
- p99
- 9.59s
- Throughput
- 70 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
parasail/fp8fp8 - Xiaomi99.85%Healthy24h 99.67%
- 7d
- 99.04%◆ (best)
- 30d
- 99.04%◆ (best)
- p50
- 2.70s
- p99
- 28.6s
- Throughput
- 36 tps
- Context
- 1.0M
- $/Mtok
- $0.140 / $0.280
xiaomi/fp8fp8 - Novita99.84%Healthy24h 97.85%
- 7d
- 98.59%
- 30d
- 98.59%
- p50
- 2.22s
- p99
- 14.9s
- Throughput
- 39 tps
- Context
- 1.0M
- $/Mtok
- $0.168 / $0.336
novita/fp8fp8 - DeepInfra98.70%Healthy24h 98.56%
- 7d
- 95.55%
- 30d
- 95.55%
- p50
- 533ms
- p99
- 2.01s◆ (best)
- Throughput
- 26 tps
- Context
- 262K
- $/Mtok
- $0.400 / $2.00
deepinfra/bf16bf16 - GMICloud97.79%Degraded24h 87.12%
- 7d
- 88.28%
- 30d
- 88.28%
- p50
- 4.15s
- p99
- 8.67s
- Throughput
- 3.0 tps
- Context
- 1.1M
- $/Mtok
- $0.119 / $0.238◆ (best)
gmicloud/fp8fp8
Measured 1m 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.