Gemma 3 4B vs Granite 4.1 8B
Wins 0 of 6 areas
—
Wins 5 of 6 areas
Coding · Agents · Facts · Long documents · Following instructions
Granite 4.1 8B is the stronger all-rounder.
Scores updated · 27 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- AgentsCarrying out multi-step tasks on its own02Granite 4.1 8B2 of 2 tests
- FactsGetting facts right instead of making them up02Granite 4.1 8B2 of 2 tests
- CodingWriting and fixing software01Granite 4.1 8B1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts01Granite 4.1 8B1 of 1 test
- Following instructionsDoing exactly what it is asked01Granite 4.1 8B1 of 1 test
- ReasoningHard problems that need careful thinking11Even1 each · 1 tie
Long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Granite 4.1 8B pulls ahead
- Code for real scientific research problemsSciCode+14.5points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+14.2points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+11.1points ahead
Where Gemma 3 4B pulls ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+1.5points ahead
Every test, side by side
All 27 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGranite 4.1 8B
- SciCodeGranite 4.1 8B by 14.57.321.8+14.5
- Terminal-Bench Hardtie0.80tie
AgentsGranite 4.1 8B
- τ-Bench V3 · BankingGranite 4.1 8B by 2.70.43.1+2.7
- GDPValGranite 4.1 8B by 2.202.2+2.2
ReasoningEven
- GPQA DiamondGranite 4.1 8B by 14.229.143.3+14.2
- Humanity's Last ExamGemma 3 4B by 1.55.33.8+1.5
- CritPttie00tie
FactsGranite 4.1 8B
- AA-Omniscience · Non-hallucinationGranite 4.1 8B by 11.11.812.9+11.1
- AA-Omniscience · AccuracyGranite 4.1 8B by 4.67.712.3+4.6
Following instructionsGranite 4.1 8B
- IFBenchGranite 4.1 8B by 10.328.338.6+10.3
Other results16 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- Arena HardGranite 4.1 8B by 37.231.869+37.2
- MBPPGranite 4.1 8B by 24.163.287.3+24.1
- τ²-Bench Telecom (AA run)Granite 4.1 8B by 22.8527.8+22.8
- AA-OmniscienceGranite 4.1 8B by 18.8-82.9-64.1+18.8
- HumanEval+Granite 4.1 8B by 17.162.879.9+17.1
- MMLUGranite 4.1 8B by 15.458.473.8+15.4
- MMMLUGranite 4.1 8B by 14.750.164.8+14.7
- HumanEvalGranite 4.1 8B by 14.171.385.4+14.1
- MMLU-ProGranite 4.1 8B by 12.443.656+12.4
- AA Agentic IndexGranite 4.1 8B by 91.710.7+9
- BBHGranite 4.1 8B by 8.372.280.5+8.3
- Artificial Analysis Coding IndexGranite 4.1 8B by 6.82.79.5+6.8
- MGSMGemma 3 4B by 587.382.3+5
- GSM8KGranite 4.1 8B by 3.389.292.5+3.3
- AA IntelligenceGranite 4.1 8B by 1.84.86.6+1.8
- SimpleQAtie44.8tie
Questions people ask
Which is better, Gemma 3 4B or Granite 4.1 8B?
Granite 4.1 8B wins five of the six areas where both have results: coding, agents, facts, long documents and following instructions. Gemma 3 4B wins none. They are level on reasoning.
Which is better for coding?
Granite 4.1 8B. It wins 1 of the 2 coding tests both models report; Gemma 3 4B wins none, and 1 is a tie.
Which is cheaper?
Gemma 3 4B costs $0.05 per million input tokens and $0.10 per million output tokens; Granite 4.1 8B costs $0.05 and $0.10. For a million tokens read plus a million written, they cost about the same.
How do you compare the two?
We use the 27 benchmark tests both models have published scores on. The verdict counts the 11 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.