Skip to content
ProviderBench home

Same-model provider benchmark

DeepInfra vs Friendli: LLM provider comparison

Compare DeepInfra and Friendli on 5 exact shared text models. ProviderBench keeps speed, price, and catalog coverage separate so naturally faster model catalogs cannot distort the result.

DeepInfra

87 indexed models

Headquarters
United States
Server regions
N/a
Model types
67 text, 15 embeddings, 5 speech

Friendli

5 indexed models

Headquarters
United States
Server regions
N/a
Model types
5 text

At a glance

Metric winners

There is no overall score. Each winner answers one specific question using only directly comparable data.

Fastest on shared models

Friendli

2.62× typical advantage

5 exact models with complete recent speed data

Lowest token cost

DeepInfra

$0.9027 / 1M

5 exact models using a 1K-input/500-output mix

Most models available

DeepInfra

87 models

Complete catalog coverage across all indexed modalities

Shared-model benchmark summary

MetricDeepInfraFriendli
Typical 500-token response26.72 s5.41 s
Typical response start0.73 s0.57 s
Typical output speed19.0 tok/s112.0 tok/s
Blended token price$0.9027 / 1M$1.3267 / 1M
Typical route uptime98.78%98.77%

Visual comparison

Price and performance charts

DeepInfraFriendli
500-token response by shared model

Estimated seconds using recent median response-start and output-speed data. Lower is better.

Friendli has the lower typical same-model response ratio across 5 measured models.

Blended token price by shared model

USD per 1 million tokens using a 1,000-input/500-output mix. Lower is better.

DeepInfra has the lower typical price ratio across 5 priced shared models.

Shared text models

5 exact models · newest first

ModelDeepInfraFriendli
GLM 5.2Winner · Friendli

z-ai/glm-5.2

DeepInfra
Blended price
$1.62 / 1M
500-token response
42.37 s
Response starts
0.70 s
Output speed
12.0 tok/s
Route uptime
N/a
Context
1M
Route
deepinfra/fp4
Friendli
Blended price
$2.4 / 1M
500-token response
4.64 s
Response starts
0.57 s
Output speed
123.0 tok/s
Route uptime
99.88%
Context
1M
Route
friendli
GLM 5.1Winner · Friendli

z-ai/glm-5.1

DeepInfra
Blended price
$1.8667 / 1M
500-token response
18.81 s
Response starts
1.57 s
Output speed
29.0 tok/s
Route uptime
99.92%
Context
202.8K
Route
deepinfra/fp4
Friendli
Blended price
$2.4 / 1M
500-token response
5.41 s
Response starts
0.95 s
Output speed
112.0 tok/s
Route uptime
97.46%
Context
202.8K
Route
friendli
MiniMax M2.5Winner · Friendli

minimax/minimax-m2.5

DeepInfra
Blended price
$0.4833 / 1M
500-token response
11.21 s
Response starts
1.01 s
Output speed
49.0 tok/s
Route uptime
100.00%
Context
196.6K
Route
deepinfra/fp8
Friendli
Blended price
$0.6 / 1M
500-token response
4.57 s
Response starts
0.57 s
Output speed
125.0 tok/s
Route uptime
N/a
Context
196.6K
Route
friendli
DeepSeek V3.2Winner · DeepInfra

deepseek/deepseek-v3.2

DeepInfra
Blended price
$0.3 / 1M
500-token response
42.39 s
Response starts
0.73 s
Output speed
12.0 tok/s
Route uptime
83.26%
Context
163.8K
Route
deepinfra/fp4
Friendli
Blended price
$0.8333 / 1M
500-token response
16.17 s
Response starts
1.02 s
Output speed
33.0 tok/s
Route uptime
99.39%
Context
163.8K
Route
friendli
Qwen3 235B A22B Instruct 2507Winner · DeepInfra

qwen/qwen3-235b-a22b-2507

DeepInfra
Blended price
$0.2433 / 1M
500-token response
26.72 s
Response starts
0.40 s
Output speed
19.0 tok/s
Route uptime
98.78%
Context
262.1K
Route
deepinfra/fp8
Friendli
Blended price
$0.4 / 1M
500-token response
17.66 s
Response starts
0.42 s
Output speed
29.0 tok/s
Route uptime
98.77%
Context
262.1K
Route
friendli

A per-model winner combines blended price and estimated 500-token response time with equal proportional weight. Ties and rows missing either measurement receive no badge. Comparison data calculated . Values use one deterministic route per provider and model; missing measurements remain visible as N/a.

DeepInfra vs Friendli analysis

How DeepInfra and Friendli compare for AI inference

DeepInfra and Friendli share 5 indexed text models, including GLM 5.2, GLM 5.1, MiniMax M2.5, DeepSeek V3.2, Qwen3 235B A22B Instruct 2507. Friendli has the stronger typical response-time result on the directly measured set.

Same-model speed evidence

5 shared models currently have complete response-start and output-speed measurements on both providers. The speed comparison uses per-model ratios before taking the median, so naturally faster model catalogs do not improve the result.

Token pricing on one workload

5 shared models have complete input and output prices on both providers. Prices use the same 1,000-input/500-output-token mix and are normalized to one million tokens for readability.

Catalog and deployment differences

DeepInfra has 87 indexed models and lists no published server regions; Friendli has 5 models and lists no regions. Verify data residency, privacy terms, limits, and production latency directly before choosing.

Related same-model benchmarks

Compare with other providers

Explore qualified alternatives with the most shared measured models. Recommendations include comparisons for both DeepInfra and Friendli.

DeepInfra vs NovitaAI

45 shared model benchmarks

Friendli vs AtlasCloud

5 shared model benchmarks

DeepInfra vs Venice

26 shared model benchmarks

Friendli vs StreamLake

5 shared model benchmarks

DeepInfra vs SiliconFlow

23 shared model benchmarks

Friendli vs DigitalOcean

4 shared model benchmarks