🛡️

Provider Trust

i

Which providers are reliable — not which models are strong.

Official list API-measured TKX probe · soon Self-reported
Provider Trust
#Provider Models▾ Prices▾ Cheapest on▾ Uptime▾ Latency▾ Precision Evidence
1
requesty
requesty
508
508418———API-measured
2
burncloud
burncloud
371
371281———API-measured
3
Amazon Bedrock
official · openrouter
328
371321100%1332ms—API-measured
4
Fireworks AI
official · openrouter
286
296228100%994ms—API-measured
5
Azure
official · openrouter
191
219133100%1886ms—API-measured
6
Novita
hf · novita-ai · official · openrouter
158
29371100%1096msFP4API-measured
7
DeepInfra
hf · official · openrouter
146
23674100%513msFP4API-measured
8
Google
official · openrouter
129
168106100%1779ms—API-measured
9
94
9426———API-measured
10
OpenAI
official · openrouter
89
9363100%2100ms—API-measured
11
Together AI
official · openrouter
86
9425100%595ms—API-measured
12
kyma
official
83
8311———Official list
13
aihubmix
official
72
7237———Official list
14
Databricks
official
67
6767———Official list
15
Nebius
official · openrouter
58
6621100%601msFP4API-measured
16
Mistral AI
official · openrouter
52
5543100%841msNVFP4API-measured
17
Oracle Cloud
official
41
4141———Official list
18
Replicate
official
40
4017———Official list
19
xAI
official
36
3634———Official list
20
Baseten
hf · official · openrouter
35
628100%427msFP4API-measured
21
Alibaba
official · openrouter
34
4812100%1218msFP8API-measured
22
Cloudflare
official · openrouter
33
379100%807msFP8API-measured
23
WandB
official
33
336———Official list
24
GMI Cloud
official · openrouter
32
338100%2073msFP8API-measured
25
IBM watsonx
official
30
3025———Official list