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
Novitanovita/bf16bf16	Healthy	99.88%◆ (best)	96.80%	92.31%	97.32%	1.22s	4.35s	41 tps	12K	$0.135 / $0.400
CoreWeavecoreweave/fp16fp16	Healthy	99.84%	99.68%	99.73%	99.83%	1.56s	6.67s	65 tps	128K	$0.710 / $0.710
Groqgroq	Healthy	99.83%	99.85%	99.82%◆ (best)	99.78%	379ms◆ (best)	1.48s	257 tps◆ (best)	131K	$0.590 / $0.790
Parasailparasail/fp8fp8	Healthy	99.72%	99.48%	98.88%	99.34%	1.55s	7.42s	56 tps	131K	$0.220 / $0.500
AkashMLakashml/fp8fp8	Healthy	99.60%	99.62%	99.53%	99.57%	3.35s	11.7s	24 tps	131K	$0.200 / $0.520
DeepInfradeepinfra/turbofp8	Healthy	99.34%	99.40%	98.30%	98.02%	718ms	20.5s	10 tps	131K	$0.100 / $0.320◆ (best)
Cloudflarecloudflare/fp8fp8	Healthy	98.81%	99.01%	98.11%	97.75%	1.23s	7.53s	34 tps	24K	$0.293 / $2.25
SambaNovasambanova-turbo	Healthy	98.57%	98.49%	99.11%	98.84%	1.06s	4.51s	118 tps	131K	$0.450 / $0.900
Togethertogether	Healthy	96.92%	96.63%	98.61%	97.97%	1.21s	6.50s	50 tps	131K	$1.04 / $1.04
Crusoecrusoe/bf16bf16	No data	—	—	—	99.90%	—	—	—	131K	$0.250 / $0.750
Googlegoogle-vertex	No data	—	—	—	99.13%	1.49s	1.49s	118 tps	128K	$0.720 / $0.720
Googlegoogle-vertex/us-central1	No data	—	—	—	99.99%◆ (best)	1.10s	1.10s◆ (best)	115 tps	128K	$0.720 / $0.720
Nebiusnebius/fp8fp8	No data	—	—	—	53.95%	—	—	—	131K	$0.130 / $0.400
PrimeIntellectprimeintellect	No data	—	—	99.65%	99.65%	—	—	—	131K	$1.04 / $2.25

Novita99.88%◆ (best)
Healthy24h 96.80%
7d
92.31%
30d
97.32%
p50
1.22s
p99
4.35s
Throughput
41 tps
Context
12K
$/Mtok
$0.135 / $0.400
novita/bf16bf16
CoreWeave99.84%
Healthy24h 99.68%
7d
99.73%
30d
99.83%
p50
1.56s
p99
6.67s
Throughput
65 tps
Context
128K
$/Mtok
$0.710 / $0.710
coreweave/fp16fp16
Groq99.83%
Healthy24h 99.85%
7d
99.82%◆ (best)
30d
99.78%
p50
379ms◆ (best)
p99
1.48s
Throughput
257 tps◆ (best)
Context
131K
$/Mtok
$0.590 / $0.790
groq
Parasail99.72%
Healthy24h 99.48%
7d
98.88%
30d
99.34%
p50
1.55s
p99
7.42s
Throughput
56 tps
Context
131K
$/Mtok
$0.220 / $0.500
parasail/fp8fp8
AkashML99.60%
Healthy24h 99.62%
7d
99.53%
30d
99.57%
p50
3.35s
p99
11.7s
Throughput
24 tps
Context
131K
$/Mtok
$0.200 / $0.520
akashml/fp8fp8
DeepInfra99.34%
Healthy24h 99.40%
7d
98.30%
30d
98.02%
p50
718ms
p99
20.5s
Throughput
10 tps
Context
131K
$/Mtok
$0.100 / $0.320◆ (best)
deepinfra/turbofp8
Cloudflare98.81%
Healthy24h 99.01%
7d
98.11%
30d
97.75%
p50
1.23s
p99
7.53s
Throughput
34 tps
Context
24K
$/Mtok
$0.293 / $2.25
cloudflare/fp8fp8
SambaNova98.57%
Healthy24h 98.49%
7d
99.11%
30d
98.84%
p50
1.06s
p99
4.51s
Throughput
118 tps
Context
131K
$/Mtok
$0.450 / $0.900
sambanova-turbo
Together96.92%
Healthy24h 96.63%
7d
98.61%
30d
97.97%
p50
1.21s
p99
6.50s
Throughput
50 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
1.49s
p99
1.49s
Throughput
118 tps
Context
128K
$/Mtok
$0.720 / $0.720
google-vertex
Google—
No data24h —
7d
—
30d
99.99%◆ (best)
p50
1.10s
p99
1.10s◆ (best)
Throughput
115 tps
Context
128K
$/Mtok
$0.720 / $0.720
google-vertex/us-central1
Nebius—
No data24h —
7d
—
30d
53.95%
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 1m 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.