Gemma 4 31B: 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
Crusoecrusoe/bf16bf16	Healthy	100%◆ (best)	99.43%	98.56%	95.61%	607ms	3.64s	23 tps	262K	$0.140 / $0.400
ModelRunmodelrun/fp4fp4	Healthy	100%◆ (best)	99.02%	99.69%◆ (best)	99.61%◆ (best)	166ms◆ (best)	3.13s◆ (best)	137 tps◆ (best)	262K	$0.750 / $1.00
SiliconFlowsiliconflow/fp8fp8	Healthy	100%◆ (best)	77.56%	94.70%	90.62%	1.32s	6.17s	44 tps	262K	$0.750 / $1.00
CoreWeavecoreweave/fp4fp4	Healthy	99.93%	99.25%	99.43%	98.41%	453ms	9.52s	40 tps	262K	$0.100 / $0.340
Friendlifriendli	Healthy	99.89%	99.60%	99.46%	98.97%	942ms	8.05s	62 tps	262K	$0.140 / $0.400
Novitanovita/bf16bf16	Down	99.69%	82.09%	82.98%	88.42%	892ms	43.8s	21 tps	262K	$0.140 / $0.400
Parasailparasail/fp8fp8	Healthy	99.50%	99.33%	99.04%	99.06%	2.03s	57.1s	17 tps	262K	$0.150 / $0.400
DeepInfradeepinfra/fp8fp8	Healthy	99.44%	98.84%	98.07%	97.57%	1.79s	10.7s	19 tps	262K	$0.200 / $0.400
DeepInfradeepinfra/turbofp4	Healthy	99.11%	99.43%	97.74%	98.72%	608ms	5.91s	27 tps	262K	$0.090 / $0.340
Venicevenice/fp4fp4	Healthy	98.91%	99.51%	97.04%	98.07%	646ms	4.39s	25 tps	256K	$0.120 / $0.360
Chuteschutes/fp4fp4	Down	70.92%	94.39%	83.95%	86.55%	2.29s	21.1s	25 tps	131K	$0.120 / $0.370
Cerebrascerebras/fp16fp16	No data	—	—	—	72.37%	—	—	—	131K	$0.990 / $1.49
CoreWeavecoreweave/bf16bf16	No data	—	—	—	—	—	—	—	262K	$0.100 / $0.340
Crusoecrusoe	No data	—	—	—	94.84%	—	—	—	262K	$0.140 / $0.400
DeepInfradeepinfra/ultrafp8	No data	—	—	78.88%	82.89%	—	—	—	131K	$0.270 / $0.760
DekaLLMdekallm	No data	—	—	95.29%	95.29%	—	—	—	262K	$0.100 / $0.330
Io Netio-net	No data	—	99.47%	98.76%	98.78%	699ms	8.30s	36 tps	262K	$0.361 / $1.09
OpenInferenceopen-inference/bf16bf16	No data	—	—	—	—	—	—	—	262K	$0.070 / $0.350◆ (best)
Phalaphala	No data	—	—	—	—	—	—	—	262K	$0.150 / $0.460
Rekareka	No data	—	—	99.49%	98.23%	—	—	—	262K	$0.080 / $0.300
SambaNovasambanova	No data	—	98.60%	97.98%	95.02%	2.08s	11.5s	91 tps	262K	$0.380 / $1.15
Togethertogether	No data	—	—	—	90.79%	—	—	—	262K	$0.390 / $0.970
Venicevenice/bf16bf16	No data	—	—	—	98.37%	—	—	—	256K	$0.120 / $0.360

Crusoe100%◆ (best)
Healthy24h 99.43%
7d
98.56%
30d
95.61%
p50
607ms
p99
3.64s
Throughput
23 tps
Context
262K
$/Mtok
$0.140 / $0.400
crusoe/bf16bf16
ModelRun100%◆ (best)
Healthy24h 99.02%
7d
99.69%◆ (best)
30d
99.61%◆ (best)
p50
166ms◆ (best)
p99
3.13s◆ (best)
Throughput
137 tps◆ (best)
Context
262K
$/Mtok
$0.750 / $1.00
modelrun/fp4fp4
SiliconFlow100%◆ (best)
Healthy24h 77.56%
7d
94.70%
30d
90.62%
p50
1.32s
p99
6.17s
Throughput
44 tps
Context
262K
$/Mtok
$0.750 / $1.00
siliconflow/fp8fp8
CoreWeave99.93%
Healthy24h 99.25%
7d
99.43%
30d
98.41%
p50
453ms
p99
9.52s
Throughput
40 tps
Context
262K
$/Mtok
$0.100 / $0.340
coreweave/fp4fp4
Friendli99.89%
Healthy24h 99.60%
7d
99.46%
30d
98.97%
p50
942ms
p99
8.05s
Throughput
62 tps
Context
262K
$/Mtok
$0.140 / $0.400
friendli
Novita99.69%
Down24h 82.09%
7d
82.98%
30d
88.42%
p50
892ms
p99
43.8s
Throughput
21 tps
Context
262K
$/Mtok
$0.140 / $0.400
novita/bf16bf16
Parasail99.50%
Healthy24h 99.33%
7d
99.04%
30d
99.06%
p50
2.03s
p99
57.1s
Throughput
17 tps
Context
262K
$/Mtok
$0.150 / $0.400
parasail/fp8fp8
DeepInfra99.44%
Healthy24h 98.84%
7d
98.07%
30d
97.57%
p50
1.79s
p99
10.7s
Throughput
19 tps
Context
262K
$/Mtok
$0.200 / $0.400
deepinfra/fp8fp8
DeepInfra99.11%
Healthy24h 99.43%
7d
97.74%
30d
98.72%
p50
608ms
p99
5.91s
Throughput
27 tps
Context
262K
$/Mtok
$0.090 / $0.340
deepinfra/turbofp4
Venice98.91%
Healthy24h 99.51%
7d
97.04%
30d
98.07%
p50
646ms
p99
4.39s
Throughput
25 tps
Context
256K
$/Mtok
$0.120 / $0.360
venice/fp4fp4
Chutes70.92%
Down24h 94.39%
7d
83.95%
30d
86.55%
p50
2.29s
p99
21.1s
Throughput
25 tps
Context
131K
$/Mtok
$0.120 / $0.370
chutes/fp4fp4
Cerebras—
No data24h —
7d
—
30d
72.37%
p50
—
p99
—
Throughput
—
Context
131K
$/Mtok
$0.990 / $1.49
cerebras/fp16fp16
CoreWeave—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.100 / $0.340
coreweave/bf16bf16
Crusoe—
No data24h —
7d
—
30d
94.84%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.140 / $0.400
crusoe
DeepInfra—
No data24h —
7d
78.88%
30d
82.89%
p50
—
p99
—
Throughput
—
Context
131K
$/Mtok
$0.270 / $0.760
deepinfra/ultrafp8
DekaLLM—
No data24h —
7d
95.29%
30d
95.29%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.100 / $0.330
dekallm
Io Net—
No data24h 99.47%
7d
98.76%
30d
98.78%
p50
699ms
p99
8.30s
Throughput
36 tps
Context
262K
$/Mtok
$0.361 / $1.09
io-net
OpenInference—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.070 / $0.350◆ (best)
open-inference/bf16bf16
Phala—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.150 / $0.460
phala
Reka—
No data24h —
7d
99.49%
30d
98.23%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.080 / $0.300
reka
SambaNova—
No data24h 98.60%
7d
97.98%
30d
95.02%
p50
2.08s
p99
11.5s
Throughput
91 tps
Context
262K
$/Mtok
$0.380 / $1.15
sambanova
Together—
No data24h —
7d
—
30d
90.79%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.390 / $0.970
together
Venice—
No data24h —
7d
—
30d
98.37%
p50
—
p99
—
Throughput
—
Context
256K
$/Mtok
$0.120 / $0.360
venice/bf16bf16

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.

Throughput

Median tokens per second.