Gemma 4 31B

google/gemma-4-31b-itInputOutput

Compare its 23 providers Incident RSSCompare with Gemma 4 26B A4B , Gemma 4 26B A4B (free), Gemma 4 31B (free)

CompareGemma 4 31B

	State
DekaLLMdekallm	Healthy	100%	97.31%	94.55%	94.55%	876ms	9.60s	28 tps	262K	$0.100 / $0.330
Io Netio-net	Healthy	100%	99.08%	98.42%	98.53%	864ms	5.08s	35 tps	262K	$0.380 / $1.15
ModelRunmodelrun/fp4fp4	Healthy	100%	99.72%	99.78%	99.59%	74ms	1.12s	147 tps	262K	$0.750 / $1.00
Parasailparasail/fp8fp8	Healthy	100%	99.31%	98.92%	99.06%	690ms	8.32s	23 tps	262K	$0.150 / $0.400
Friendlifriendli	Healthy	99.87%	99.08%	98.53%	98.95%	199ms	3.67s	97 tps	262K	$0.140 / $0.400
Venicevenice/fp4fp4	Healthy	99.54%	98.98%	99.17%	98.75%	485ms	3.72s	29 tps	256K	$0.120 / $0.360
CoreWeavecoreweave/fp4fp4	Healthy	99.31%	99.55%	99.02%	98.42%	443ms	6.98s	35 tps	262K	$0.100 / $0.340
DeepInfradeepinfra/turbofp4	Healthy	99.21%	98.92%	99.18%	99.11%	469ms	5.53s	34 tps	262K	$0.090 / $0.340
DeepInfradeepinfra/fp8fp8	Healthy	98.74%	97.80%	97.97%	97.11%	2.15s	14.3s	17 tps	262K	$0.150 / $0.400
Chuteschutes/fp4fp4	Healthy	98.00%	96.20%	81.82%	86.52%	1.82s	13.0s	26 tps	131K	$0.120 / $0.370
Novitanovita/bf16bf16	Healthy	97.30%	81.77%	82.48%	89.06%	937ms	20.8s	26 tps	262K	$0.140 / $0.400
SiliconFlowsiliconflow/fp8fp8	Healthy	95.85%	83.87%	94.40%	90.65%	1.40s	8.94s	40 tps	262K	$0.750 / $1.00
Cerebrascerebras/fp16fp16	No data	—	—	—	83.32%	—	—	—	131K	$0.990 / $1.49
CoreWeavecoreweave/bf16bf16	No data	—	—	—	—	—	—	—	262K	$0.100 / $0.340
Crusoecrusoe	No data	—	—	—	95.23%	—	—	—	262K	$0.140 / $0.400
Crusoecrusoe/bf16bf16	Degraded	—	98.92%	98.83%	95.16%	486ms	14.6s	18 tps	262K	$0.140 / $0.400
DeepInfradeepinfra/ultrafp8	No data	—	—	89.63%	84.22%	—	—	—	131K	$0.270 / $0.760
OpenInferenceopen-inference/bf16bf16	No data	—	—	—	—	—	—	—	262K	$0.070 / $0.350
Phalaphala	No data	—	—	—	—	—	—	—	262K	$0.150 / $0.460
Rekareka	No data	—	—	99.62%	98.23%	—	—	—	262K	$0.080 / $0.300
SambaNovasambanova	No data	—	97.65%	97.44%	94.94%	2.44s	8.12s	96 tps	131K	$0.380 / $1.15
Togethertogether	No data	—	—	—	92.43%	—	—	—	262K	$0.390 / $0.970
Venicevenice/bf16bf16	No data	—	—	—	98.60%	—	—	—	256K	$0.120 / $0.360

DekaLLM100%
Healthy24h 97.31%
7d
94.55%
30d
94.55%
p50
876ms
p99
9.60s
Throughput
28 tps
Context
262K
$/Mtok
$0.100 / $0.330
dekallm
Io Net100%
Healthy24h 99.08%
7d
98.42%
30d
98.53%
p50
864ms
p99
5.08s
Throughput
35 tps
Context
262K
$/Mtok
$0.380 / $1.15
io-net
ModelRun100%
Healthy24h 99.72%
7d
99.78%
30d
99.59%
p50
74ms
p99
1.12s
Throughput
147 tps
Context
262K
$/Mtok
$0.750 / $1.00
modelrun/fp4fp4
Parasail100%
Healthy24h 99.31%
7d
98.92%
30d
99.06%
p50
690ms
p99
8.32s
Throughput
23 tps
Context
262K
$/Mtok
$0.150 / $0.400
parasail/fp8fp8
Friendli99.87%
Healthy24h 99.08%
7d
98.53%
30d
98.95%
p50
199ms
p99
3.67s
Throughput
97 tps
Context
262K
$/Mtok
$0.140 / $0.400
friendli
Venice99.54%
Healthy24h 98.98%
7d
99.17%
30d
98.75%
p50
485ms
p99
3.72s
Throughput
29 tps
Context
256K
$/Mtok
$0.120 / $0.360
venice/fp4fp4
CoreWeave99.31%
Healthy24h 99.55%
7d
99.02%
30d
98.42%
p50
443ms
p99
6.98s
Throughput
35 tps
Context
262K
$/Mtok
$0.100 / $0.340
coreweave/fp4fp4
DeepInfra99.21%
Healthy24h 98.92%
7d
99.18%
30d
99.11%
p50
469ms
p99
5.53s
Throughput
34 tps
Context
262K
$/Mtok
$0.090 / $0.340
deepinfra/turbofp4
DeepInfra98.74%
Healthy24h 97.80%
7d
97.97%
30d
97.11%
p50
2.15s
p99
14.3s
Throughput
17 tps
Context
262K
$/Mtok
$0.150 / $0.400
deepinfra/fp8fp8
Chutes98.00%
Healthy24h 96.20%
7d
81.82%
30d
86.52%
p50
1.82s
p99
13.0s
Throughput
26 tps
Context
131K
$/Mtok
$0.120 / $0.370
chutes/fp4fp4
Novita97.30%
Healthy24h 81.77%
7d
82.48%
30d
89.06%
p50
937ms
p99
20.8s
Throughput
26 tps
Context
262K
$/Mtok
$0.140 / $0.400
novita/bf16bf16
SiliconFlow95.85%
Healthy24h 83.87%
7d
94.40%
30d
90.65%
p50
1.40s
p99
8.94s
Throughput
40 tps
Context
262K
$/Mtok
$0.750 / $1.00
siliconflow/fp8fp8
Cerebras—
No data24h —
7d
—
30d
83.32%
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
95.23%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.140 / $0.400
crusoe
Crusoe—
Degraded24h 98.92%
7d
98.83%
30d
95.16%
p50
486ms
p99
14.6s
Throughput
18 tps
Context
262K
$/Mtok
$0.140 / $0.400
crusoe/bf16bf16
DeepInfra—
No data24h —
7d
89.63%
30d
84.22%
p50
—
p99
—
Throughput
—
Context
131K
$/Mtok
$0.270 / $0.760
deepinfra/ultrafp8
OpenInference—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.070 / $0.350
open-inference/bf16bf16
Phala—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.150 / $0.460
phala
Reka—
No data24h —
7d
99.62%
30d
98.23%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.080 / $0.300
reka
SambaNova—
No data24h 97.65%
7d
97.44%
30d
94.94%
p50
2.44s
p99
8.12s
Throughput
96 tps
Context
131K
$/Mtok
$0.380 / $1.15
sambanova
Together—
No data24h —
7d
—
30d
92.43%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.390 / $0.970
together
Venice—
No data24h —
7d
—
30d
98.60%
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.

History

9 endpoints have no chartable data in this range.

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.

Daily availability

One row per endpoint · one dot per day · last 90 days.

worst hour ≥ 95%90–95%< 90%no data

Colour by each endpoint's worst hour of each UTC day · 23 endpoints · 47 days with data, 304 endpoint-days below 90%.

Price changes

Gemma 4 31B ModelRunmodelrun/fp4$0.750 / $1.00cache read $0.750 → $0.2002026-10-02 15:20 UTC
Gemma 4 31B DeepInfradeepinfra/fp8$0.130 → $0.150 / $0.380 → $0.4002026-10-01 20:20 UTC
Gemma 4 31B DeepInfradeepinfra/fp8$0.150 → $0.130 / $0.400 → $0.3802026-10-01 20:15 UTC
Gemma 4 31B DeepInfradeepinfra/fp8$0.130 → $0.150 / $0.380 → $0.4002026-10-01 20:05 UTC
Gemma 4 31B DekaLLMdekallm$0.060 → $0.100 / $0.330cache read — → $0.0502026-09-30 15:31 UTC
Gemma 4 31B DekaLLMdekallm$0.100 → $0.060 / $0.330cache read $0.050 → —2026-09-30 15:26 UTC
Gemma 4 31B DekaLLMdekallm$0.060 → $0.100 / $0.330cache read — → $0.0502026-09-30 15:06 UTC
Gemma 4 31B Rekareka$0.090 → $0.080 / $0.340 → $0.3002026-09-26 10:45 UTC
Gemma 4 31B Io Netio-net$0.380 / $1.15cache read — → $0.1902026-09-25 15:55 UTC
Gemma 4 31B SiliconFlowsiliconflow/fp8$0.130 → $0.750 / $0.400 → $1.00cache read — → $0.2502026-09-10 05:40 UTC
Gemma 4 31B OpenInferenceopen-inference/bf16$0.080 → $0.070 / $0.350cache read $0.010 → $0.1002026-08-29 01:40 UTC
Gemma 4 31B Togethertogether$0.280 → $0.390 / $0.860 → $0.9702026-08-27 21:55 UTC
Gemma 4 31B Togethertogether$0.390 → $0.280 / $0.970 → $0.8602026-08-19 02:50 UTC
Gemma 4 31B Togethertogether$0.280 → $0.390 / $0.860 → $0.9702026-08-19 02:45 UTC