GLM 4.6: 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Healthy | 100%◆ (best) | 99.90% | 99.25%◆ (best) | 99.25%◆ (best) | 504ms◆ (best) | 3.70s◆ (best) | 38 tps | 203K | $0.500 / $2.00 | |
| Healthy | 100%◆ (best) | 97.44% | 96.82% | 96.82% | 997ms | 14.8s | 35 tps | 205K | $0.550 / $2.20 | |
| Healthy | 100%◆ (best) | 99.35% | 98.86% | 98.86% | 1.16s | 17.0s | 18 tps | 198K | $0.430 / $1.75◆ (best) | |
| Healthy | 97.14% | 96.03% | 91.87% | 91.87% | 859ms | 27.3s | 37 tps | 203K | $0.600 / $2.20 | |
| Down | — | 93.55% | 97.13% | 97.13% | 611ms | 7.10s | 45 tps◆ (best) | 203K | $0.600 / $2.20 |
- DeepInfra100%◆ (best)Healthy24h 99.90%
- 7d
- 99.25%◆ (best)
- 30d
- 99.25%◆ (best)
- p50
- 504ms◆ (best)
- p99
- 3.70s◆ (best)
- Throughput
- 38 tps
- Context
- 203K
- $/Mtok
- $0.500 / $2.00
deepinfra/fp4fp4 - Novita100%◆ (best)Healthy24h 97.44%
- 7d
- 96.82%
- 30d
- 96.82%
- p50
- 997ms
- p99
- 14.8s
- Throughput
- 35 tps
- Context
- 205K
- $/Mtok
- $0.550 / $2.20
novita/bf16bf16 - Venice100%◆ (best)Healthy24h 99.35%
- 7d
- 98.86%
- 30d
- 98.86%
- p50
- 1.16s
- p99
- 17.0s
- Throughput
- 18 tps
- Context
- 198K
- $/Mtok
- $0.430 / $1.75◆ (best)
venice/fp4fp4 - Z.AI97.14%Healthy24h 96.03%
- 7d
- 91.87%
- 30d
- 91.87%
- p50
- 859ms
- p99
- 27.3s
- Throughput
- 37 tps
- Context
- 203K
- $/Mtok
- $0.600 / $2.20
z-ai/fp4fp4 - Down24h 93.55%
- 7d
- 97.13%
- 30d
- 97.13%
- p50
- 611ms
- p99
- 7.10s
- Throughput
- 45 tps◆ (best)
- Context
- 203K
- $/Mtok
- $0.600 / $2.20
atlas-cloud/fp8fp8
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.