GPT-5.4 Mini: 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Azureazure | Healthy | 100%◆ (best) | 99.97% | 99.98% | 99.98% | 867ms | 5.15s | 77 tps | 400K | $0.750 / $4.50 |
OpenAIopenai | Healthy | 100%◆ (best) | 99.99% | 99.92% | 99.63% | 828ms | 4.13s | 118 tps◆ (best) | 400K | $0.750 / $4.50 |
Azureazure/us | No data | — | — | 100%◆ (best) | 99.01% | — | — | — | 400K | $0.825 / $4.95 |
OpenAIopenai/fast | No data | — | 100% | 99.98% | 99.98%◆ (best) | 957ms | 3.00s | 100 tps | 400K | $1.50 / $9.00 |
OpenAIopenai/flex | No data | — | 100% | 100% | 99.00% | 795ms◆ (best) | 1.06s◆ (best) | 84 tps | 400K | $0.375 / $2.25◆ (best) |
OpenAIopenai/priority | No data | — | — | — | — | — | — | — | 400K | $1.50 / $9.00 |
- Azure100%◆ (best)Healthy24h 99.97%
- 7d
- 99.98%
- 30d
- 99.98%
- p50
- 867ms
- p99
- 5.15s
- Throughput
- 77 tps
- Context
- 400K
- $/Mtok
- $0.750 / $4.50
azure - OpenAI100%◆ (best)Healthy24h 99.99%
- 7d
- 99.92%
- 30d
- 99.63%
- p50
- 828ms
- p99
- 4.13s
- Throughput
- 118 tps◆ (best)
- Context
- 400K
- $/Mtok
- $0.750 / $4.50
openai - No data24h —
- 7d
- 100%◆ (best)
- 30d
- 99.01%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 400K
- $/Mtok
- $0.825 / $4.95
azure/us - No data24h 100%
- 7d
- 99.98%
- 30d
- 99.98%◆ (best)
- p50
- 957ms
- p99
- 3.00s
- Throughput
- 100 tps
- Context
- 400K
- $/Mtok
- $1.50 / $9.00
openai/fast - No data24h 100%
- 7d
- 100%
- 30d
- 99.00%
- p50
- 795ms◆ (best)
- p99
- 1.06s◆ (best)
- Throughput
- 84 tps
- Context
- 400K
- $/Mtok
- $0.375 / $2.25◆ (best)
openai/flex - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 400K
- $/Mtok
- $1.50 / $9.00
openai/priority
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, averaged over the hour.
Throughput
Median tokens per second.