What is the best IBM model?

Granite 4.2 8B has the highest verification score of the IBM models we tested: 36.2% (29 of 80 correct) in our 2026-10-27 run.

What is the fastest IBM model?

Granite 4.2 8B is the fastest IBM model we tested, at a median 60 output tokens per second.

What is the cheapest IBM model?

Granite 4.0 Micro has the lowest billed cost per question of the IBM models we tested: $0.000051.

Which IBM models do you test?

Granite 4.2 8B, Granite 4.0 Micro.

Provider

IBM AI models

2 models tested, compared with 76 models from other labs.

Best verification score36.2Granite 4.2 8B · #60 of 78 Best output speed60Granite 4.2 8B · #53 of 78 Best time to first token0.6 sGranite 4.2 8B · #25 of 78 Best cost per question$0.000051Granite 4.0 Micro · #1 of 78

IBM#60Granite 4.2 8B36.2% verification score60 tok/s0.6 s first token$0.000077 / question

IBM#69Granite 4.0 Micro25.0% verification score38 tok/s0.7 s first token$0.000051 / question

Score vs. cost per question

IBM models in color, every other model in grey

Output speed vs. cost per question

IBM models in color, every other model in grey

Verification score

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

60 more models · see the full leaderboard

Output speed

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

60 more models · see the full leaderboard

Time to first token

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

62 more models · see the full leaderboard

Cost per question

IBM 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

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

62 more models · see the full leaderboard

Context window

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

66 more models · see the full leaderboard

All IBM results

Leaderboard
#
1	Granite 4.2 8B IBM · served by DeepInfra	36.2	29 / 80	60	0.6 s	$0.000077	$0.06 / $0.25	131k
2	Granite 4.0 Micro IBM · served by Cloudflare	25.0	20 / 80	38	0.7 s	$0.000051	$0.017 / $0.112	131k

IBM: quick answers

What is the best IBM model?: Granite 4.2 8B has the highest verification score of the IBM models we tested: 36.2% (29 of 80 correct) in our 2026-10-27 run.
What is the fastest IBM model?: Granite 4.2 8B is the fastest IBM model we tested, at a median 60 output tokens per second.
What is the cheapest IBM model?: Granite 4.0 Micro has the lowest billed cost per question of the IBM models we tested: $0.000051.
Which IBM models do you test?: Granite 4.2 8B, Granite 4.0 Micro.

Provider

IBM AI models

2 models tested, compared with 76 models from other labs.

IBM#60Granite 4.2 8B36.2% verification score60 tok/s0.6 s first token$0.000077 / question

IBM#69Granite 4.0 Micro25.0% verification score38 tok/s0.7 s first token$0.000051 / question

Score vs. cost per question

IBM models in color, every other model in grey

Output speed vs. cost per question

IBM models in color, every other model in grey

Verification score

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

60 more models · see the full leaderboard

Output speed

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

60 more models · see the full leaderboard

Time to first token

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

62 more models · see the full leaderboard

Cost per question

IBM 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

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

62 more models · see the full leaderboard

Context window

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

66 more models · see the full leaderboard

All IBM results

Leaderboard
#
1	Granite 4.2 8B IBM · served by DeepInfra	36.2	29 / 80	60	0.6 s	$0.000077	$0.06 / $0.25	131k
2	Granite 4.0 Micro IBM · served by Cloudflare	25.0	20 / 80	38	0.7 s	$0.000051	$0.017 / $0.112	131k

IBM: quick answers

What is the best IBM model?: Granite 4.2 8B has the highest verification score of the IBM models we tested: 36.2% (29 of 80 correct) in our 2026-10-27 run.
What is the fastest IBM model?: Granite 4.2 8B is the fastest IBM model we tested, at a median 60 output tokens per second.
What is the cheapest IBM model?: Granite 4.0 Micro has the lowest billed cost per question of the IBM models we tested: $0.000051.
Which IBM models do you test?: Granite 4.2 8B, Granite 4.0 Micro.