Qwen3.6 35B A3B: 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
AkashMLakashml/fp8fp8	Healthy	100%◆ (best)	100%	99.76%	99.80%	518ms◆ (best)	2.45s◆ (best)	97 tps	262K	$0.100 / $0.900
AtlasCloudatlas-cloud/fp8fp8	Healthy	100%◆ (best)	99.74%	98.27%	98.68%	907ms	5.89s	84 tps	262K	$0.186 / $1.11
CoreWeavecoreweave/fp8fp8	Healthy	100%◆ (best)	100%	99.90%	98.99%	620ms	5.26s	155 tps◆ (best)	262K	$0.250 / $1.25
Darkbloomdarkbloom/fp4fp4	Healthy	100%◆ (best)	100%	99.91%	99.84%	2.23s	12.4s	76 tps	262K	$0.050 / $0.700◆ (best)
DeepInfradeepinfra/fp8fp8	Healthy	100%◆ (best)	100%	99.03%	96.14%	1.13s	24.4s	40 tps	262K	$0.100 / $0.950
DekaLLMdekallm	Healthy	100%◆ (best)	99.86%	99.63%	99.24%	683ms	9.68s	107 tps	262K	$0.100 / $1.00
Parasailparasail/fp8fp8	Healthy	100%◆ (best)	99.65%	99.86%	99.76%	554ms	3.15s	116 tps	262K	$0.150 / $1.00
Phalaphala	Healthy	100%◆ (best)	97.59%	99.12%	99.41%	1.05s	16.7s	49 tps	262K	$0.200 / $1.27
SiliconFlowsiliconflow/fp8fp8	Healthy	100%◆ (best)	99.63%	98.88%	98.62%	1.83s	12.7s	67 tps	262K	$0.240 / $1.80
Venicevenice/fp8fp8	Healthy	100%◆ (best)	99.71%	99.76%	99.84%	781ms	6.35s	101 tps	256K	$0.100 / $1.00
Darkbloomdarkbloom	No data	—	—	—	—	—	—	—	262K	$0.080 / $0.750
Io Netio-net/fp8fp8	No data	—	—	—	95.18%	—	—	—	262K	$0.220 / $1.20
Rekareka	No data	—	—	100%◆ (best)	99.91%◆ (best)	—	—	—	262K	$0.100 / $1.00

AkashML100%◆ (best)
Healthy24h 100%
7d
99.76%
30d
99.80%
p50
518ms◆ (best)
p99
2.45s◆ (best)
Throughput
97 tps
Context
262K
$/Mtok
$0.100 / $0.900
akashml/fp8fp8
AtlasCloud100%◆ (best)
Healthy24h 99.74%
7d
98.27%
30d
98.68%
p50
907ms
p99
5.89s
Throughput
84 tps
Context
262K
$/Mtok
$0.186 / $1.11
atlas-cloud/fp8fp8
CoreWeave100%◆ (best)
Healthy24h 100%
7d
99.90%
30d
98.99%
p50
620ms
p99
5.26s
Throughput
155 tps◆ (best)
Context
262K
$/Mtok
$0.250 / $1.25
coreweave/fp8fp8
Darkbloom100%◆ (best)
Healthy24h 100%
7d
99.91%
30d
99.84%
p50
2.23s
p99
12.4s
Throughput
76 tps
Context
262K
$/Mtok
$0.050 / $0.700◆ (best)
darkbloom/fp4fp4
DeepInfra100%◆ (best)
Healthy24h 100%
7d
99.03%
30d
96.14%
p50
1.13s
p99
24.4s
Throughput
40 tps
Context
262K
$/Mtok
$0.100 / $0.950
deepinfra/fp8fp8
DekaLLM100%◆ (best)
Healthy24h 99.86%
7d
99.63%
30d
99.24%
p50
683ms
p99
9.68s
Throughput
107 tps
Context
262K
$/Mtok
$0.100 / $1.00
dekallm
Parasail100%◆ (best)
Healthy24h 99.65%
7d
99.86%
30d
99.76%
p50
554ms
p99
3.15s
Throughput
116 tps
Context
262K
$/Mtok
$0.150 / $1.00
parasail/fp8fp8
Phala100%◆ (best)
Healthy24h 97.59%
7d
99.12%
30d
99.41%
p50
1.05s
p99
16.7s
Throughput
49 tps
Context
262K
$/Mtok
$0.200 / $1.27
phala
SiliconFlow100%◆ (best)
Healthy24h 99.63%
7d
98.88%
30d
98.62%
p50
1.83s
p99
12.7s
Throughput
67 tps
Context
262K
$/Mtok
$0.240 / $1.80
siliconflow/fp8fp8
Venice100%◆ (best)
Healthy24h 99.71%
7d
99.76%
30d
99.84%
p50
781ms
p99
6.35s
Throughput
101 tps
Context
256K
$/Mtok
$0.100 / $1.00
venice/fp8fp8
Darkbloom—
No data24h —
7d
—
30d
—
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.080 / $0.750
darkbloom
Io Net—
No data24h —
7d
—
30d
95.18%
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.220 / $1.20
io-net/fp8fp8
Reka—
No data24h —
7d
100%◆ (best)
30d
99.91%◆ (best)
p50
—
p99
—
Throughput
—
Context
262K
$/Mtok
$0.100 / $1.00
reka

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, averaged over the hour.

Throughput

Median tokens per second.