GLM 5.3 Flash: 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
BaseTenbaseten/fp8fp8	Healthy	100%◆ (best)	99.19%	98.68%	98.68%	826ms◆ (best)	5.68s◆ (best)	152 tps◆ (best)	1.0M	$0.150 / $0.500
Io Netio-net/fp8fp8	Healthy	100%◆ (best)	99.90%	100%◆ (best)	100%◆ (best)	2.47s	26.4s	13 tps	262K	$0.150 / $0.500
Parasailparasail/fp8fp8	Healthy	100%◆ (best)	99.46%	99.13%	99.13%	919ms	51.6s	48 tps	1.0M	$0.150 / $0.500
Novitanovita/fp8fp8	Healthy	99.48%	99.47%	99.40%	99.40%	1.73s	77.9s	42 tps	1.0M	$0.075 / $0.250◆ (best)
Cloudflarecloudflare	Healthy	98.27%	98.23%	77.70%	77.70%	3.39s	21.6s	45 tps	1.3M	$0.150 / $0.500
Z.AIz-ai/fp8fp8	Healthy	97.85%	99.34%	99.21%	99.21%	5.53s	79.8s	30 tps	1.0M	$0.075 / $0.250◆ (best)
GMICloudgmicloud/fp8fp8	Healthy	96.78%	98.16%	98.65%	98.65%	3.08s	29.7s	39 tps	1.0M	$0.075 / $0.250◆ (best)
Rekareka/fp8fp8	Degraded	90.67%	98.59%	—	—	2.59s	60.4s	39 tps	262K	$0.075 / $0.250◆ (best)
Venicevenice	Down	53.79%	49.03%	69.93%	69.93%	5.25s	132s	14 tps	1.0M	$0.094 / $0.313
DeepInfradeepinfra/fp8fp8	No data	—	99.02%	96.96%	96.96%	135s	265s	6.0 tps	1.0M	$0.150 / $0.500
Modalmodal/fp8fp8	No data	—	92.46%	93.11%	93.11%	1.52s	8.44s	78 tps	1.0M	$0.150 / $0.500
Togethertogether	No data	—	—	—	—	—	—	—	1.0M	$0.150 / $0.500

BaseTen100%◆ (best)
Healthy24h 99.19%
7d
98.68%
30d
98.68%
p50
826ms◆ (best)
p99
5.68s◆ (best)
Throughput
152 tps◆ (best)
Context
1.0M
$/Mtok
$0.150 / $0.500
baseten/fp8fp8
Io Net100%◆ (best)
Healthy24h 99.90%
7d
100%◆ (best)
30d
100%◆ (best)
p50
2.47s
p99
26.4s
Throughput
13 tps
Context
262K
$/Mtok
$0.150 / $0.500
io-net/fp8fp8
Parasail100%◆ (best)
Healthy24h 99.46%
7d
99.13%
30d
99.13%
p50
919ms
p99
51.6s
Throughput
48 tps
Context
1.0M
$/Mtok
$0.150 / $0.500
parasail/fp8fp8
Novita99.48%
Healthy24h 99.47%
7d
99.40%
30d
99.40%
p50
1.73s
p99
77.9s
Throughput
42 tps
Context
1.0M
$/Mtok
$0.075 / $0.250◆ (best)
novita/fp8fp8
Cloudflare98.27%
Healthy24h 98.23%
7d
77.70%
30d
77.70%
p50
3.39s
p99
21.6s
Throughput
45 tps
Context
1.3M
$/Mtok
$0.150 / $0.500
cloudflare
Z.AI97.85%
Healthy24h 99.34%
7d
99.21%
30d
99.21%
p50
5.53s
p99
79.8s
Throughput
30 tps
Context
1.0M
$/Mtok
$0.075 / $0.250◆ (best)
z-ai/fp8fp8
GMICloud96.78%
Healthy24h 98.16%
7d
98.65%
30d
98.65%
p50
3.08s
p99
29.7s
Throughput
39 tps
Context
1.0M
$/Mtok
$0.075 / $0.250◆ (best)
gmicloud/fp8fp8
Reka90.67%
Degraded24h 98.59%
7d
—
30d
—
p50
2.59s
p99
60.4s
Throughput
39 tps
Context
262K
$/Mtok
$0.075 / $0.250◆ (best)
reka/fp8fp8
Venice53.79%
Down24h 49.03%
7d
69.93%
30d
69.93%
p50
5.25s
p99
132s
Throughput
14 tps
Context
1.0M
$/Mtok
$0.094 / $0.313
venice
DeepInfra—
No data24h 99.02%
7d
96.96%
30d
96.96%
p50
135s
p99
265s
Throughput
6.0 tps
Context
1.0M
$/Mtok
$0.150 / $0.500
deepinfra/fp8fp8
Modal—
No data24h 92.46%
7d
93.11%
30d
93.11%
p50
1.52s
p99
8.44s
Throughput
78 tps
Context
1.0M
$/Mtok
$0.150 / $0.500
modal/fp8fp8
Together—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
1.0M
$/Mtok
$0.150 / $0.500
together

Measured 5m 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.