Microsoft
Microsoft is an AI model provider.Tokenando tracks 10 Microsoft models, with input pricing from $0.040/M and an average blended cost of $0.242/M. Its flagship model is Phi 4.
Phi small language models from Microsoft Research — efficient reasoning at sizes that fit on-device or on a single GPU.
MODELS TRACKED
10
1 category
FLAGSHIP
Phi 4
text->text
MIN INPUT
$0.040/M
cheapest model in family
AVG BLENDED
$0.242/M
across 10 priced models
MAX CONTEXT
128,000
largest window in family
Efficient
10 models
Phi-3 Smallprofile
Live API · 128,000 ctx
in $0.150/Mout $0.600/M
7B small model · manual-seed
Phi 4profile
text->text · 16,384 ctx
in $0.070/Mout $0.140/M
tokenizer: Other · cron:openrouter
Phi-4 Miniprofile
Live API · 128,000 ctx
in $0.040/Mout $0.160/M
3.8B mini reasoning · manual-seed
Phi-3.5 Miniprofile
Live API · 128,000 ctx
in $0.130/Mout $0.520/M
3.8B long-context · manual-seed
Phi-3.5 MoEprofile
Live API · 128,000 ctx
in $0.160/Mout $0.640/M
16x3.8B mixture · manual-seed
Phi-3 Mediumprofile
Live API · 128,000 ctx
in $0.170/Mout $0.680/M
14B medium · manual-seed
Phi-3 Miniprofile
Live API · 128,000 ctx
in $0.130/Mout $0.520/M
3.8B mini · manual-seed
Phi-2profileDEPRECATED
Live API · 2,000 ctx
in $0.070/Mout $0.070/M
Original Phi 2.7B · manual-seed
WizardLM-2 8x22Bprofile
text->text · 65,535 ctx
in $0.620/Mout $0.620/M
tokenizer: Mistral · cron:openrouter
Phi 4 Mini Instructprofile
text->text · 128,000 ctx
in $0.080/Mout $0.350/M
tokenizer: Other · cron:openrouter
Frequently Asked Questions
How many models does Microsoft offer?
Tokenando tracks 10 Microsoft models.
How much do Microsoft models cost?
Microsoft model input pricing starts at $0.040 per million tokens, with an average blended cost of $0.242 per million across the 10 priced models we track.
What is Microsoft's flagship model?
Microsoft's flagship model is Phi 4. It is the highest-tier Microsoft model we track, with input pricing of $0.070 per million tokens.
What model categories does Microsoft cover?
Microsoft covers 1 category: efficient.