GLM 5.2: 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) | 98.75% | 99.87% | 99.87% | 1.24s | 5.64s | 57 tps | 1.0M | $2.31 / $7.26 | |
| Healthy | 100%◆ (best) | 99.17% | 99.80% | 99.80% | 1.38s | 9.41s | 42 tps | 1.0M | $1.26 / $3.96 | |
| Healthy | 100%◆ (best) | 99.40% | 99.39% | 99.39% | 1.93s | 14.5s | 85 tps | 1.0M | $2.10 / $6.60 | |
| Healthy | 100%◆ (best) | 99.53% | 99.59% | 99.59% | 1.63s | 15.8s | 69 tps | 1.0M | $1.40 / $4.40 | |
Cloudflarecloudflare/fast | Healthy | 100%◆ (best) | 99.90% | 99.93%◆ (best) | 99.93%◆ (best) | 1.36s | 28.3s | 61 tps | 262K | $2.10 / $6.60 |
| Healthy | 100%◆ (best) | 99.68% | 99.76% | 99.76% | 754ms | 3.64s◆ (best) | 51 tps | 1.0M | $1.40 / $4.40 | |
| Healthy | 100%◆ (best) | 99.88% | 99.89% | 99.89% | 660ms | 7.44s | 88 tps | 262K | $1.40 / $4.40 | |
| Healthy | 100%◆ (best) | 99.12% | 99.72% | 99.72% | 1.54s | 24.2s | 43 tps | 1.0M | $1.19 / $3.74 | |
| Healthy | 100%◆ (best) | 99.18% | 99.71% | 99.71% | 2.12s | 12.3s | 47 tps | 1.0M | $0.448 / $1.41◆ (best) | |
| Healthy | 100%◆ (best) | 98.45% | 98.84% | 98.84% | 4.90s | 9.38s | 46 tps | 1.0M | $1.26 / $3.96 | |
| Healthy | 99.97% | 99.88% | 99.63% | 99.63% | 1.19s | 9.87s | 54 tps | 1.0M | $1.40 / $4.40 | |
Friendlifriendli | Healthy | 99.94% | 99.72% | 99.25% | 99.25% | 615ms | 9.41s | 93 tps | 1.0M | $1.40 / $4.40 |
| Healthy | 99.93% | 95.27% | 99.47% | 99.47% | 1.32s | 13.9s | 30 tps | 1.0M | $0.742 / $2.33 | |
| Healthy | 99.86% | 99.88% | 99.83% | 99.83% | 1.27s | 9.78s | 54 tps | 1.0M | $0.462 / $1.45 | |
| Healthy | 99.86% | 99.51% | 99.30% | 99.30% | 1.55s | 24.0s | 29 tps | 1.0M | $0.750 / $2.40 | |
| Healthy | 99.72% | 99.59% | 99.63% | 99.63% | 1.67s | 25.4s | 29 tps | 1.0M | $0.500 / $3.15 | |
Fireworksfireworks/fast-us | Healthy | 99.51% | 96.33% | 96.83% | 96.83% | 1.62s | 14.0s | 103 tps◆ (best) | 1.0M | $2.10 / $6.60 |
| Healthy | 99.30% | 96.18% | 96.78% | 96.78% | 548ms◆ (best) | 13.2s | 54 tps | 1.0M | $0.750 / $2.90 | |
Fireworksfireworks | Healthy | 99.22% | 98.35% | 98.72% | 98.72% | 1.64s | 18.6s | 47 tps | 1.0M | $1.40 / $4.40 |
| Healthy | 99.18% | 95.97% | 99.71% | 99.71% | 1.47s | 14.6s | 43 tps | 1.0M | $1.26 / $3.00 | |
DigitalOceandigitalocean | Healthy | 99.13% | 99.32% | 99.12% | 99.12% | 1.10s | 18.7s | 36 tps | 262K | $0.700 / $2.20 |
| Healthy | 98.94% | 99.53% | 99.42% | 99.42% | 1.60s | 21.3s | 40 tps | 1.0M | $0.741 / $2.33 | |
| Healthy | 98.06% | 98.28% | 98.57% | 98.57% | 1.32s | 11.0s | 45 tps | 1.0M | $1.40 / $4.40 | |
| Healthy | 97.92% | 98.67% | 99.59% | 99.59% | 842ms | 9.80s | 49 tps | 1.0M | $0.966 / $3.04 | |
Fireworksfireworks/fast | Degraded | 94.87% | 96.95% | 97.49% | 97.49% | 1.32s | 14.7s | 89 tps | 1.0M | $2.10 / $6.60 |
| Degraded | 92.63% | 92.35% | 95.93% | 95.93% | 1.76s | 23.5s | 9.0 tps | 1.0M | $0.842 / $3.14 | |
| Degraded | 91.76% | 97.51% | 98.44% | 98.44% | 3.91s | 19.3s | 34 tps | 1.0M | $0.684 / $2.28 | |
Togethertogether | Down | 85.97% | 90.86% | 95.21% | 95.21% | 756ms | 13.7s | 60 tps | 512K | $1.40 / $4.40 |
| Down | 66.67% | 90.65% | 92.73% | 92.73% | 959ms | 27.1s | 48 tps | 203K | $0.600 / $2.00 | |
| No data | — | — | 97.12% | 97.12% | — | — | — | 97K | $0.770 / $2.42 | |
Cloudflarecloudflare | No data | — | 100% | 99.86% | 99.86% | 1.74s | 9.81s | 49 tps | 262K | $1.40 / $4.40 |
| No data | — | 99.78% | 99.80% | 99.80% | 753ms | 7.33s | 62 tps | 262K | $0.760 / $2.42 |
- Alibaba100%◆ (best)Healthy24h 98.75%
- 7d
- 99.87%
- 30d
- 99.87%
- p50
- 1.24s
- p99
- 5.64s
- Throughput
- 57 tps
- Context
- 1.0M
- $/Mtok
- $2.31 / $7.26
alibaba/fastfp8 - AtlasCloud100%◆ (best)Healthy24h 99.17%
- 7d
- 99.80%
- 30d
- 99.80%
- p50
- 1.38s
- p99
- 9.41s
- Throughput
- 42 tps
- Context
- 1.0M
- $/Mtok
- $1.26 / $3.96
atlas-cloud/fp8fp8 - BaseTen100%◆ (best)Healthy24h 99.40%
- 7d
- 99.39%
- 30d
- 99.39%
- p50
- 1.93s
- p99
- 14.5s
- Throughput
- 85 tps
- Context
- 1.0M
- $/Mtok
- $2.10 / $6.60
baseten/fastfp8 - BaseTen100%◆ (best)Healthy24h 99.53%
- 7d
- 99.59%
- 30d
- 99.59%
- p50
- 1.63s
- p99
- 15.8s
- Throughput
- 69 tps
- Context
- 1.0M
- $/Mtok
- $1.40 / $4.40
baseten/fp8fp8 - Cloudflare100%◆ (best)Healthy24h 99.90%
- 7d
- 99.93%◆ (best)
- 30d
- 99.93%◆ (best)
- p50
- 1.36s
- p99
- 28.3s
- Throughput
- 61 tps
- Context
- 262K
- $/Mtok
- $2.10 / $6.60
cloudflare/fast - Crusoe100%◆ (best)Healthy24h 99.68%
- 7d
- 99.76%
- 30d
- 99.76%
- p50
- 754ms
- p99
- 3.64s◆ (best)
- Throughput
- 51 tps
- Context
- 1.0M
- $/Mtok
- $1.40 / $4.40
crusoe/fp8fp8 - Parasail100%◆ (best)Healthy24h 99.88%
- 7d
- 99.89%
- 30d
- 99.89%
- p50
- 660ms
- p99
- 7.44s
- Throughput
- 88 tps
- Context
- 262K
- $/Mtok
- $1.40 / $4.40
parasail/fp4fp4 - SiliconFlow100%◆ (best)Healthy24h 99.12%
- 7d
- 99.72%
- 30d
- 99.72%
- p50
- 1.54s
- p99
- 24.2s
- Throughput
- 43 tps
- Context
- 1.0M
- $/Mtok
- $1.19 / $3.74
siliconflow/fp8fp8 - StreamLake100%◆ (best)Healthy24h 99.18%
- 7d
- 99.71%
- 30d
- 99.71%
- p50
- 2.12s
- p99
- 12.3s
- Throughput
- 47 tps
- Context
- 1.0M
- $/Mtok
- $0.448 / $1.41◆ (best)
streamlake/fp8fp8 - Wafer100%◆ (best)Healthy24h 98.45%
- 7d
- 98.84%
- 30d
- 98.84%
- p50
- 4.90s
- p99
- 9.38s
- Throughput
- 46 tps
- Context
- 1.0M
- $/Mtok
- $1.26 / $3.96
waferfp4 - Z.AI99.97%Healthy24h 99.88%
- 7d
- 99.63%
- 30d
- 99.63%
- p50
- 1.19s
- p99
- 9.87s
- Throughput
- 54 tps
- Context
- 1.0M
- $/Mtok
- $1.40 / $4.40
z-ai/fp8fp8 - Friendli99.94%Healthy24h 99.72%
- 7d
- 99.25%
- 30d
- 99.25%
- p50
- 615ms
- p99
- 9.41s
- Throughput
- 93 tps
- Context
- 1.0M
- $/Mtok
- $1.40 / $4.40
friendli - GMICloud99.93%Healthy24h 95.27%
- 7d
- 99.47%
- 30d
- 99.47%
- p50
- 1.32s
- p99
- 13.9s
- Throughput
- 30 tps
- Context
- 1.0M
- $/Mtok
- $0.742 / $2.33
gmicloud/fp8fp8 - Baidu99.86%Healthy24h 99.88%
- 7d
- 99.83%
- 30d
- 99.83%
- p50
- 1.27s
- p99
- 9.78s
- Throughput
- 54 tps
- Context
- 1.0M
- $/Mtok
- $0.462 / $1.45
baidu/fp8fp8 - DeepInfra99.86%Healthy24h 99.51%
- 7d
- 99.30%
- 30d
- 99.30%
- p50
- 1.55s
- p99
- 24.0s
- Throughput
- 29 tps
- Context
- 1.0M
- $/Mtok
- $0.750 / $2.40
deepinfra/fp4fp4 - Sail Research99.72%Healthy24h 99.59%
- 7d
- 99.63%
- 30d
- 99.63%
- p50
- 1.67s
- p99
- 25.4s
- Throughput
- 29 tps
- Context
- 1.0M
- $/Mtok
- $0.500 / $3.15
sail-research/fp8fp8 - Fireworks99.51%Healthy24h 96.33%
- 7d
- 96.83%
- 30d
- 96.83%
- p50
- 1.62s
- p99
- 14.0s
- Throughput
- 103 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $2.10 / $6.60
fireworks/fast-us - Inceptron99.30%Healthy24h 96.18%
- 7d
- 96.78%
- 30d
- 96.78%
- p50
- 548ms◆ (best)
- p99
- 13.2s
- Throughput
- 54 tps
- Context
- 1.0M
- $/Mtok
- $0.750 / $2.90
inceptron/fp4fp4 - Fireworks99.22%Healthy24h 98.35%
- 7d
- 98.72%
- 30d
- 98.72%
- p50
- 1.64s
- p99
- 18.6s
- Throughput
- 47 tps
- Context
- 1.0M
- $/Mtok
- $1.40 / $4.40
fireworks - Phala99.18%Healthy24h 95.97%
- 7d
- 99.71%
- 30d
- 99.71%
- p50
- 1.47s
- p99
- 14.6s
- Throughput
- 43 tps
- Context
- 1.0M
- $/Mtok
- $1.26 / $3.00
phala/fp8fp8 - DigitalOcean99.13%Healthy24h 99.32%
- 7d
- 99.12%
- 30d
- 99.12%
- p50
- 1.10s
- p99
- 18.7s
- Throughput
- 36 tps
- Context
- 262K
- $/Mtok
- $0.700 / $2.20
digitalocean - Novita98.94%Healthy24h 99.53%
- 7d
- 99.42%
- 30d
- 99.42%
- p50
- 1.60s
- p99
- 21.3s
- Throughput
- 40 tps
- Context
- 1.0M
- $/Mtok
- $0.741 / $2.33
novita/fp8fp8 - Venice98.06%Healthy24h 98.28%
- 7d
- 98.57%
- 30d
- 98.57%
- p50
- 1.32s
- p99
- 11.0s
- Throughput
- 45 tps
- Context
- 1.0M
- $/Mtok
- $1.40 / $4.40
venice/fp8fp8 - Alibaba97.92%Healthy24h 98.67%
- 7d
- 99.59%
- 30d
- 99.59%
- p50
- 842ms
- p99
- 9.80s
- Throughput
- 49 tps
- Context
- 1.0M
- $/Mtok
- $0.966 / $3.04
alibaba/fp8fp8 - Fireworks94.87%Degraded24h 96.95%
- 7d
- 97.49%
- 30d
- 97.49%
- p50
- 1.32s
- p99
- 14.7s
- Throughput
- 89 tps
- Context
- 1.0M
- $/Mtok
- $2.10 / $6.60
fireworks/fast - Morph92.63%Degraded24h 92.35%
- 7d
- 95.93%
- 30d
- 95.93%
- p50
- 1.76s
- p99
- 23.5s
- Throughput
- 9.0 tps
- Context
- 1.0M
- $/Mtok
- $0.842 / $3.14
morph/fp4fp4 - Decart91.76%Degraded24h 97.51%
- 7d
- 98.44%
- 30d
- 98.44%
- p50
- 3.91s
- p99
- 19.3s
- Throughput
- 34 tps
- Context
- 1.0M
- $/Mtok
- $0.684 / $2.28
decart/fp4fp4 - Together85.97%Down24h 90.86%
- 7d
- 95.21%
- 30d
- 95.21%
- p50
- 756ms
- p99
- 13.7s
- Throughput
- 60 tps
- Context
- 512K
- $/Mtok
- $1.40 / $4.40
together - Ambient66.67%Down24h 90.65%
- 7d
- 92.73%
- 30d
- 92.73%
- p50
- 959ms
- p99
- 27.1s
- Throughput
- 48 tps
- Context
- 203K
- $/Mtok
- $0.600 / $2.00
ambient/fp8fp8 - No data24h —
- 7d
- 97.12%
- 30d
- 97.12%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 97K
- $/Mtok
- $0.770 / $2.42
akashml/fp8fp8 - No data24h 100%
- 7d
- 99.86%
- 30d
- 99.86%
- p50
- 1.74s
- p99
- 9.81s
- Throughput
- 49 tps
- Context
- 262K
- $/Mtok
- $1.40 / $4.40
cloudflare - No data24h 99.78%
- 7d
- 99.80%
- 30d
- 99.80%
- p50
- 753ms
- p99
- 7.33s
- Throughput
- 62 tps
- Context
- 262K
- $/Mtok
- $0.760 / $2.42
coreweave/fp4fp4
Measured 3m 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, averaged over the hour.
Throughput
Median tokens per second.