Gemma 4 31B vs GPT-5.5
Wins 0 of 8 areas
—
Wins 6 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents
GPT-5.5 is the stronger all-rounder.Gemma 4 31B is cheaper.
Scores updated · 40 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software05GPT-5.55 of 5 tests
- AgentsCarrying out multi-step tasks on its own05GPT-5.55 of 5 tests
- ReasoningHard problems that need careful thinking03GPT-5.53 of 3 tests
- Images and chartsUnderstanding pictures, charts and video03GPT-5.53 of 3 tests
- MathCompetition and research-level math02GPT-5.52 of 2 tests
- Long documentsFinding answers in very long texts01GPT-5.51 of 1 test
- FactsGetting facts right instead of making them up22Even2 each
- Following instructionsDoing exactly what it is asked00Even0 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.
Gemma 4 31B costs 98% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where GPT-5.5 pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+52.6points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+40.9points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+38points ahead
Where Gemma 4 31B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+4points ahead
Every test, side by side
All 40 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.5
- Terminal-Bench 2.1GPT-5.5 by 40.943.484.3+40.9
- Terminal-Bench HardGPT-5.5 by 24.236.460.6+24.2
- SWE-bench ProGPT-5.5 by 22.935.758.6+22.9
- LMArena · WebDevGPT-5.5 by 148 rating points13651513+148 rating
- SciCodeGPT-5.5 by 10.345.555.8+10.3
AgentsGPT-5.5
- GDPValGPT-5.5 by 36.66.142.7+36.6
- τ-Bench V3 · BankingGPT-5.5 by 24.214.839+24.2
- MCP AtlasGPT-5.5 by 18.157.275.3+18.1
- Terminal-Bench 4.0GPT-5.5 by 14.6014.6+14.6
- AA IT-Bench SREGPT-5.5 by 8.537.345.8+8.5
ReasoningGPT-5.5
- CritPtGPT-5.5 by 25.71.427.1+25.7
- Humanity's Last ExamGPT-5.5 by 22.223.645.8+22.2
- GPQA DiamondGPT-5.5 by 7.885.793.5+7.8
FactsEven
- SimpleQA VerifiedGPT-5.5 by 52.610.463+52.6
- AA-Omniscience · AccuracyGPT-5.5 by 382058+38
- AA-Omniscience · Non-hallucinationGemma 4 31B by 41511+4
- Vectara HHEM hallucination ratelower is betterGemma 4 31B by 1.97.49.3+1.9
Images and chartsGPT-5.5
- MMMU-ProGPT-5.5 by 6.573.479.9+6.5
- MathVistaGPT-5.5 by 4.979.384.2+4.9
- LMArena · VisionGPT-5.5 by 15 rating points12771292+15 rating
MathGPT-5.5
- HMMT Feb. 2026GPT-5.5 by 21.377.298.5+21.3
- AIME 2026GPT-5.5 by 10.889.2100+10.8
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.
- AA-OmniscienceGPT-5.5 by 68.4-47.920.5+68.4
- Terminal-Bench 2.0GPT-5.5 by 39.842.982.7+39.8
- NL2RepoGPT-5.5 by 35.215.550.7+35.2
- Tool-DecathlonGPT-5.5 by 34.421.255.6+34.4
- τ²-Bench Telecom (AA run)GPT-5.5 by 3459.993.9+34
- Artificial Analysis Coding IndexGPT-5.5 by 31.543.474.9+31.5
- AA Agentic IndexGPT-5.5 by 30.66.737.3+30.6
- HLE (with tools)GPT-5.5 by 25.726.552.2+25.7
- AA IntelligenceGPT-5.5 by 23.714.738.4+23.7
- RealWorldQAGPT-5.5 by 9.972.382.2+9.9
- HMMT Nov. 2025GPT-5.5 by 987.596.5+9
- SuperGPQAGPT-5.5 by 765.772.7+7
- SimpleVQAGPT-5.5 by 5.752.958.6+5.7
- vectara_avg_summary_lengthGPT-5.5 by 53.8 rating points75.8129.6+53.8 rating
- vectara_factual_consistencyGemma 4 31B by 1.992.690.7+1.9
- vectara_answer_ratetie100100tie
Questions people ask
Which is better, Gemma 4 31B or GPT-5.5?
GPT-5.5 wins six of the eight areas we test: coding, agents, reasoning, images and charts, math and long documents. Gemma 4 31B wins none, but costs 98% less. They are level on facts and following instructions.
Which is better for coding?
GPT-5.5. It wins 5 of the 5 coding tests both models report; Gemma 4 31B wins none.
Which is cheaper?
Gemma 4 31B costs $0.17 per million input tokens and $0.40 per million output tokens; GPT-5.5 costs $5.00 and $30.00. That makes Gemma 4 31B about 98% cheaper for the same work.
How do you compare the two?
We use the 40 benchmark tests both models have published scores on. The verdict counts the 24 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.