Granite 4.1 3B
Apr 2026
Granite 4.1 3BCompare inon
Benchmarks of Granite 4.1 3B13
BenchmarkByScoreRank
Overall1
Knowledge4
Science2
Coding1
Agents1
Documents1
Legacy · saturated or no longer reported3
No benchmark matches
—per task?
self-host onlyprice
—output speed?
131Kcontext
Ranking?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-6 Lunalow×3.7—$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinking×3.5—$0.30 / $1.20Memory (4-bit): 1.7 GB → 4.5 GBBetterGranite 4.2 3BthinkingOpen+54%—$0.006Context: 131K → 128K+1
BetterGemma 4 E4BthinkingOpen+52%—$0.008Context: 131K → 128K+1BetterGPT-5.6 Lunalow×3.6—$0.01ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×3.7—$0.005Function calling only with none · Chat Completions+3Granite 4.2 3BthinkingOpen+54%—$0.006Context: 131K → 128K+1Gemma 4 E4BthinkingOpen+52%—$0.008Context: 131K → 128K+1GPT-5.6 Lunalow×3.6—$0.01GPT OSS 20BhighOpen+53%—$0.011Memory (4-bit): 1.7 GB → 13 GBModelScoreCostPer taskWhat changes
IK2 Horizon 7BthinkingOpen×3.5—$0.30 / $1.20Memory (4-bit): 1.7 GB → 4.5 GBIK2 Horizon 3.7BthinkingOpen×2.7—$0.30 / $1.20Memory (4-bit): 1.7 GB → 2.5 GBGemma 4 12BthinkingOpen×2.4—$0.10 / $0.30Memory (4-bit): 1.7 GB → 6 GB
Qwen 3.5 9Bnon-thinkingOpen×2.3—$0.08 / $0.13Memory (4-bit): 1.7 GB → 4.8 GBQwen3.5 4BthinkingOpen×2.2—$0.03 / $0.15Memory (4-bit): 1.7 GB → 2.3 GBModelScoreCostPer taskWhat changes
Granite 4.2 30BthinkingOpen×2.2—$0.072Context: 131K → 128K+1Granite 4.2 8BthinkingOpen+89%—$0.024Context: 131K → 128K+1Granite 4.2 3BthinkingOpen+54%—$0.006Context: 131K → 128K+1Granite 4.1 30BOpen+27%—$0.10 / $0.25Memory (4-bit): 1.7 GB → 17 GBGranite 4.1 8Bnon-thinkingOpensame——Memory (4-bit): 1.7 GB → 4.4 GBGranite 4.0 H Smallnon-thinkingOpensame—$0.05 / $0.22Context: 131K → 128K+1Granite 4.0 H 1BOpensame——Context: 131K → 128KGranite 4.0 Micronon-thinkingOpensame—$0.017 / $0.11All IBM models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · full quality✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use
History
2 versionsBenchmarks
all categoriesChat & Knowledge
Humanity's Last Exam3.4#329
AA Omniscience Index-77#306
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈5.9#300
AA Omniscience · Accuracy9.3#286
GPQA Diamond31.4#296
Coding
Terminal-Bench 2.11.1#149
Agents & Tool Use
τ²-bench · AA19.6#196
Writing & Language
IFBench (instructions)33.7#194
FAQ
How much does Granite 4.1 3B cost?›
There is no per-unit API price: self-host only.
Is Granite 4.1 3B free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is Granite 4.1 3B, and what is it best at?›
#326 of 334 in Chat & Knowledge, #181 of 239 in Coding, #122 of 175 in Agents & Tool Use, #192 of 245 in Writing & Language.
Is Granite 4.1 3B better than Granite 4.0 1B?›
On AA Intelligence Index 4.3.2, Granite 4.1 3B scores 5.9 vs 5 — about the same.
What is the context window of Granite 4.1 3B?›
131K tokens.
When was Granite 4.1 3B released? Is there a newer version?›
Released 2026-04-29. It has been superseded by Granite 4.2 3B.
Is Granite 4.1 3B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Granite 4.1 3B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Granite 4.1 3B read images?›
No, it accepts text only.
What is Granite 4.1 3B?›
Granite 4.1 3B is an open-weights language model from IBM, released Apr 2026, the successor to Granite 4.0 1B.
What are the best alternatives to Granite 4.1 3B?›
GPT-6 Luna (better), K2 Horizon 7B (open weights).