Skip to content

Fastest First-Token Claude Sonnet 4.6 Inference Providers

Ranks providers by how quickly the first generated token appears. Providers without enough recent data remain visible after ranked rows.

Amazon Bedrock currently ranks first at 0.88 s via amazon-bedrock/global.

Provider options:
8
Ranked:
8
Metric:
Time to first token

Fastest first token endpoint ranking

Fastest First-Token Claude Sonnet 4.6 Inference Providers
RankProvider / routeTime to first tokenSpeed measurementsEndpoint details
#1Provider / route
Amazon Bedrockamazon-bedrock/global
Time to first token0.88 sSpeed measurements
500-token response
10.88 s
Response starts
0.88 s
Output speed
50.0 tok/s
Endpoint details
Input price
$3 / 1M tokens
Output price
$15 / 1M tokens
Uptime (30 min)
100.00%
Context
1M
#2Provider / route
Anthropicanthropic
Time to first token1.19 sSpeed measurements
500-token response
17.32 s
Response starts
1.19 s
Output speed
31.0 tok/s
Endpoint details
Input price
$3 / 1M tokens
Output price
$15 / 1M tokens
Uptime (30 min)
99.95%
Context
1M
#3Provider / route
Anthropicanthropic/2
Time to first token1.22 sSpeed measurements
500-token response
14.04 s
Response starts
1.22 s
Output speed
39.0 tok/s
Endpoint details
Input price
$3 / 1M tokens
Output price
$15 / 1M tokens
Uptime (30 min)
99.99%
Context
1M
#4Provider / route
Google Vertexgoogle-vertex/global
Time to first token1.26 sSpeed measurements
500-token response
14.41 s
Response starts
1.26 s
Output speed
38.0 tok/s
Endpoint details
Input price
$3 / 1M tokens
Output price
$15 / 1M tokens
Uptime (30 min)
100.00%
Context
1M
#5Provider / route
Google Vertexgoogle-vertex/europe
Time to first token1.33 sSpeed measurements
500-token response
22.17 s
Response starts
1.33 s
Output speed
24.0 tok/s
Endpoint details
Input price
$3.3 / 1M tokens
Output price
$16.5 / 1M tokens
Uptime (30 min)
100.00%
Context
1M
#6Provider / route
Google Vertexgoogle-vertex/us-east5
Time to first token1.44 sSpeed measurements
500-token response
11.65 s
Response starts
1.44 s
Output speed
49.0 tok/s
Endpoint details
Input price
$3.3 / 1M tokens
Output price
$16.5 / 1M tokens
Uptime (30 min)
Insufficient recent data
Context
1M
#7Provider / route
Amazon Bedrockamazon-bedrock/eu-west-1
Time to first token1.55 sSpeed measurements
500-token response
12.19 s
Response starts
1.55 s
Output speed
47.0 tok/s
Endpoint details
Input price
$3.3 / 1M tokens
Output price
$16.5 / 1M tokens
Uptime (30 min)
Insufficient recent data
Context
1M
#8Provider / route
Azureazure/us-east-2
Time to first token2.75 sSpeed measurements
500-token response
15.41 s
Response starts
2.75 s
Output speed
39.5 tok/s
Endpoint details
Input price
$3 / 1M tokens
Output price
$15 / 1M tokens
Uptime (30 min)
Insufficient recent data
Context
1M

Endpoint data fetched .

Claude Sonnet 4.6 endpoint guide

How to interpret the Fastest First-Token Claude Sonnet 4.6 Inference Providers

8 of 8 Claude Sonnet 4.6 endpoints currently have the published data required for this ranking. Amazon Bedrock leads at 0.88 s via amazon-bedrock/global. The table keeps unranked routes visible so missing measurements do not look like missing provider availability.

How response-start time is ranked

The first-token ranking orders endpoints by the recent median time before output begins. This is useful for interactive applications, but it should be read alongside throughput because the first provider to start is not always the first to finish a long answer.

Compare Amazon Bedrock with the next option

Anthropic currently ranks second at 1.19 s. Compare that gap with input and output price, response-start time, output speed, 30-minute uptime, context length, and the exact route before deciding whether first place is meaningful for your workload.

Use the route, not only the provider name

Claude Sonnet 4.6 has 8 published provider options, and performance data is matched to each exact OpenRouter routing tag. Different quantization, context, regional deployment, or provider configuration can change price and behavior even when the underlying model name is identical.