Llama 3.3 70B Instruct: 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
Groqgroq	Healthy	100%◆ (best)	99.86%	99.82%◆ (best)	99.78%	384ms◆ (best)	1.37s◆ (best)	254 tps◆ (best)	131K	$0.590 / $0.790
Novitanovita/bf16bf16	Healthy	99.88%	97.17%	92.31%	97.32%	1.29s	4.35s	38 tps	12K	$0.135 / $0.400
Parasailparasail/fp8fp8	Healthy	99.36%	99.49%	98.92%	99.34%	1.60s	5.74s	54 tps	131K	$0.220 / $0.500
Cloudflarecloudflare/fp8fp8	Healthy	98.58%	99.03%	98.11%	97.76%	500ms	5.94s	36 tps	24K	$0.293 / $2.25
AkashMLakashml/fp8fp8	Healthy	97.55%	99.62%	99.53%	99.57%	2.33s	11.8s	26 tps	131K	$0.200 / $0.520
DeepInfradeepinfra/turbofp8	Healthy	96.72%	99.38%	98.31%	98.02%	748ms	19.8s	10 tps	131K	$0.100 / $0.320◆ (best)
SambaNovasambanova-turbo	Healthy	95.36%	98.48%	99.11%	98.84%	967ms	4.71s	128 tps	131K	$0.450 / $0.900
CoreWeavecoreweave/fp16fp16	Down	86.93%	99.53%	99.73%	99.83%	1.62s	11.0s	48 tps	128K	$0.710 / $0.710
Togethertogether	Down	73.51%	92.71%	98.58%	97.97%	4.01s	16.8s	25 tps	131K	$1.04 / $1.04
Crusoecrusoe/bf16bf16	No data	—	—	—	99.90%	—	—	—	131K	$0.250 / $0.750
Googlegoogle-vertex	No data	—	—	—	99.13%	—	—	—	128K	$0.720 / $0.720
Googlegoogle-vertex/us-central1	No data	—	—	—	99.98%◆ (best)	—	—	—	128K	$0.720 / $0.720
Nebiusnebius/fp8fp8	No data	—	—	—	53.83%	—	—	—	131K	$0.130 / $0.400
PrimeIntellectprimeintellect	No data	—	—	99.65%	99.65%	—	—	—	131K	$1.04 / $2.25

Groq100%◆ (best)
Healthy24h 99.86%
7d
99.82%◆ (best)
30d
99.78%
p50
384ms◆ (best)
p99
1.37s◆ (best)
Throughput
254 tps◆ (best)
Context
131K
$/Mtok
$0.590 / $0.790
groq
Novita99.88%
Healthy24h 97.17%
7d
92.31%
30d
97.32%
p50
1.29s
p99
4.35s
Throughput
38 tps
Context
12K
$/Mtok
$0.135 / $0.400
novita/bf16bf16
Parasail99.36%
Healthy24h 99.49%
7d
98.92%
30d
99.34%
p50
1.60s
p99
5.74s
Throughput
54 tps
Context
131K
$/Mtok
$0.220 / $0.500
parasail/fp8fp8
Cloudflare98.58%
Healthy24h 99.03%
7d
98.11%
30d
97.76%
p50
500ms
p99
5.94s
Throughput
36 tps
Context
24K
$/Mtok
$0.293 / $2.25
cloudflare/fp8fp8
AkashML97.55%
Healthy24h 99.62%
7d
99.53%
30d
99.57%
p50
2.33s
p99
11.8s
Throughput
26 tps
Context
131K
$/Mtok
$0.200 / $0.520
akashml/fp8fp8
DeepInfra96.72%
Healthy24h 99.38%
7d
98.31%
30d
98.02%
p50
748ms
p99
19.8s
Throughput
10 tps
Context
131K
$/Mtok
$0.100 / $0.320◆ (best)
deepinfra/turbofp8
SambaNova95.36%
Healthy24h 98.48%
7d
99.11%
30d
98.84%
p50
967ms
p99
4.71s
Throughput
128 tps
Context
131K
$/Mtok
$0.450 / $0.900
sambanova-turbo
CoreWeave86.93%
Down24h 99.53%
7d
99.73%
30d
99.83%
p50
1.62s
p99
11.0s
Throughput
48 tps
Context
128K
$/Mtok
$0.710 / $0.710
coreweave/fp16fp16
Together73.51%
Down24h 92.71%
7d
98.58%
30d
97.97%
p50
4.01s
p99
16.8s
Throughput
25 tps
Context
131K
$/Mtok
$1.04 / $1.04
together
Crusoe—
No data24h —
7d
—
30d
99.90%
p50
—
p99
—
Throughput
—
Context
131K
$/Mtok
$0.250 / $0.750
crusoe/bf16bf16
Google—
No data24h —
7d
—
30d
99.13%
p50
—
p99
—
Throughput
—
Context
128K
$/Mtok
$0.720 / $0.720
google-vertex
Google—
No data24h —
7d
—
30d
99.98%◆ (best)
p50
—
p99
—
Throughput
—
Context
128K
$/Mtok
$0.720 / $0.720
google-vertex/us-central1
Nebius—
No data24h —
7d
—
30d
53.83%
p50
—
p99
—
Throughput
—
Context
131K
$/Mtok
$0.130 / $0.400
nebius/fp8fp8
PrimeIntellect—
No data24h —
7d
99.65%
30d
99.65%
p50
—
p99
—
Throughput
—
Context
131K
$/Mtok
$1.04 / $2.25
primeintellect

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.