What is the best Microsoft model?

Phi 4 has the highest verification score of the Microsoft models we tested: 41.2% (33 of 80 correct) in our 2026-10-27 run.

What is the fastest Microsoft model?

Phi 4 is the fastest Microsoft model we tested, at a median 103 output tokens per second.

What is the cheapest Microsoft model?

Phi 4 has the lowest billed cost per question of the Microsoft models we tested: $0.00012.

Provider

Microsoft AI models

1 model tested, compared with 77 models from other labs.

Best verification score41.2Phi 4 · #54 of 78 Best output speed103Phi 4 · #35 of 78 Best time to first token0.5 sPhi 4 · #9 of 78 Best cost per question$0.00012Phi 4 · #6 of 78

Microsoft#54Phi 441.2% verification score103 tok/s0.5 s first token$0.00012 / question

Score vs. cost per question

Microsoft models in color, every other model in grey

Output speed vs. cost per question

Microsoft models in color, every other model in grey

Verification score

Microsoft models highlighted against every model in this run · higher is better

61 more models · see the full leaderboard

Output speed

Microsoft models highlighted against every model in this run · higher is better

61 more models · see the full leaderboard

Time to first token

Microsoft models highlighted against every model in this run · lower is better

64 more models · see the full leaderboard

Cost per question

Microsoft models highlighted against every model in this run · lower is better

1Granite 4.0 Micro$0.000051
2gpt-oss-20b$0.000066
3Granite 4.2 8B$0.000077
4Nova Micro 1.0$0.000087
5Mercury 2.5$0.00011
6Phi 4$0.00012
7gpt-oss-120b$0.00013
8GLM 5.3 Flash$0.00015
9Ministral 3 3B 2512$0.00019
10GPT-6 Luna$0.00025

62 more models · see the full leaderboard

End-to-end response time

Microsoft models highlighted against every model in this run · lower is better

63 more models · see the full leaderboard

Context window

Microsoft models highlighted against every model in this run · higher is better

67 more models · see the full leaderboard

All Microsoft results

Leaderboard
#
1	Phi 4 Microsoft · served by DeepInfra	41.2	33 / 80	103	0.5 s	$0.00012	$0.07 / $0.14	16k

Microsoft: quick answers

What is the best Microsoft model?: Phi 4 has the highest verification score of the Microsoft models we tested: 41.2% (33 of 80 correct) in our 2026-10-27 run.
What is the fastest Microsoft model?: Phi 4 is the fastest Microsoft model we tested, at a median 103 output tokens per second.
What is the cheapest Microsoft model?: Phi 4 has the lowest billed cost per question of the Microsoft models we tested: $0.00012.
Which Microsoft models do you test?: Phi 4.

Provider

Microsoft AI models

1 model tested, compared with 77 models from other labs.

Best verification score41.2Phi 4 · #54 of 78 Best output speed103Phi 4 · #35 of 78 Best time to first token0.5 sPhi 4 · #9 of 78 Best cost per question$0.00012Phi 4 · #6 of 78

Microsoft#54Phi 441.2% verification score103 tok/s0.5 s first token$0.00012 / question

Score vs. cost per question

Microsoft models in color, every other model in grey

Output speed vs. cost per question

Microsoft models in color, every other model in grey

Verification score

Microsoft models highlighted against every model in this run · higher is better

61 more models · see the full leaderboard

Output speed

Microsoft models highlighted against every model in this run · higher is better

61 more models · see the full leaderboard

Time to first token

Microsoft models highlighted against every model in this run · lower is better

64 more models · see the full leaderboard

Cost per question

Microsoft models highlighted against every model in this run · lower is better

1Granite 4.0 Micro$0.000051
2gpt-oss-20b$0.000066
3Granite 4.2 8B$0.000077
4Nova Micro 1.0$0.000087
5Mercury 2.5$0.00011
6Phi 4$0.00012
7gpt-oss-120b$0.00013
8GLM 5.3 Flash$0.00015
9Ministral 3 3B 2512$0.00019
10GPT-6 Luna$0.00025

62 more models · see the full leaderboard

End-to-end response time

Microsoft models highlighted against every model in this run · lower is better

63 more models · see the full leaderboard

Context window

Microsoft models highlighted against every model in this run · higher is better

67 more models · see the full leaderboard

All Microsoft results

Leaderboard
#
1	Phi 4 Microsoft · served by DeepInfra	41.2	33 / 80	103	0.5 s	$0.00012	$0.07 / $0.14	16k

Microsoft: quick answers

What is the best Microsoft model?: Phi 4 has the highest verification score of the Microsoft models we tested: 41.2% (33 of 80 correct) in our 2026-10-27 run.
What is the fastest Microsoft model?: Phi 4 is the fastest Microsoft model we tested, at a median 103 output tokens per second.
What is the cheapest Microsoft model?: Phi 4 has the lowest billed cost per question of the Microsoft models we tested: $0.00012.
Which Microsoft models do you test?: Phi 4.