Gemma 4 31B vs Qwen3.7 Plus Preview
Wins 0 of 8 areas
—
Wins 8 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents · Following instructions
Qwen3.7 Plus Preview is the stronger all-rounder.Gemma 4 31B is cheaper.
Scores updated · 46 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 software06Qwen3.7 Plus Preview6 of 7 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own04Qwen3.7 Plus Preview4 of 4 tests
- ReasoningHard problems that need careful thinking03Qwen3.7 Plus Preview3 of 3 tests
- Images and chartsUnderstanding pictures, charts and video02Qwen3.7 Plus Preview2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up02Qwen3.7 Plus Preview2 of 2 tests
- MathCompetition and research-level math02Qwen3.7 Plus Preview2 of 2 tests
- Long documentsFinding answers in very long texts01Qwen3.7 Plus Preview1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.7 Plus Preview1 of 1 test
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 72% 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 Qwen3.7 Plus Preview pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+57.3points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+25.7points ahead
- Fixes real GitHub issues in many programming languagesSWE-bench Multilingual+24.1points ahead
Where Gemma 4 31B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 46 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.7 Plus Preview
- SWE-bench VerifiedQwen3.7 Plus Preview by 25.75277.7+25.7
- SWE-bench MultilingualQwen3.7 Plus Preview by 24.151.775.8+24.1
- SWE-bench ProQwen3.7 Plus Preview by 21.935.757.6+21.9
- Terminal-Bench 2.1Qwen3.7 Plus Preview by 17.643.461+17.6
- Terminal-Bench HardQwen3.7 Plus Preview by 10.636.447+10.6
- LiveCodeBench v6Qwen3.7 Plus Preview by 9.68089.6+9.6
- SciCodetie45.546.1tie
AgentsQwen3.7 Plus Preview
- MCP AtlasQwen3.7 Plus Preview by 1657.273.2+16
- GDPValQwen3.7 Plus Preview by 7.4613.4+7.4
- τ-Bench V3 · BankingQwen3.7 Plus Preview by 2.714.817.5+2.7
- Terminal-Bench 4.0Qwen3.7 Plus Preview by 101+1
ReasoningQwen3.7 Plus Preview
- Humanity's Last ExamQwen3.7 Plus Preview by 1223.635.6+12
- CritPtQwen3.7 Plus Preview by 7.71.49.1+7.7
- GPQA DiamondQwen3.7 Plus Preview by 4.385.790+4.3
FactsQwen3.7 Plus Preview
- AA-Omniscience · Non-hallucinationQwen3.7 Plus Preview by 57.31572.3+57.3
- AA-Omniscience · AccuracyQwen3.7 Plus Preview by 2.52022.5+2.5
Images and chartsQwen3.7 Plus Preview
- CharXiv (RQ)Qwen3.7 Plus Preview by 1867.985.9+18
- MMMU-ProQwen3.7 Plus Preview by 7.173.480.5+7.1
- LMArena · Visiontie12791277tie
MathQwen3.7 Plus Preview
- HMMT Feb. 2026Qwen3.7 Plus Preview by 15.777.292.9+15.7
- IMOAnswerBenchQwen3.7 Plus Preview by 11.574.586+11.5
Long documentsQwen3.7 Plus Preview
- AA-LCRQwen3.7 Plus Preview by 3.369.773+3.3
Following instructionsQwen3.7 Plus Preview
- IFBenchQwen3.7 Plus Preview by 2.475.678+2.4
Other results23 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-OmniscienceQwen3.7 Plus Preview by 49-47.91.1+49
- MCPMarkQwen3.7 Plus Preview by 40.618.158.7+40.6
- DeepPlanningQwen3.7 Plus Preview by 38.32462.3+38.3
- Claw Eval (pass@3)Qwen3.7 Plus Preview by 37.72562.7+37.7
- τ²-Bench Telecom (AA run)Qwen3.7 Plus Preview by 33.159.993+33.1
- SimpleVQAQwen3.7 Plus Preview by 28.852.981.7+28.8
- Terminal-Bench 2.0Qwen3.7 Plus Preview by 27.442.970.3+27.4
- NL2RepoQwen3.7 Plus Preview by 25.615.541.1+25.6
- QwenClawBenchQwen3.7 Plus Preview by 20.141.761.8+20.1
- RealWorldQAQwen3.7 Plus Preview by 14.672.386.9+14.6
- AA Agentic IndexQwen3.7 Plus Preview by 136.719.7+13
- Artificial Analysis Coding IndexQwen3.7 Plus Preview by 12.543.455.9+12.5
- OmniDocBench 1.5Qwen3.7 Plus Preview by 11.380.191.4+11.3
- AA IntelligenceQwen3.7 Plus Preview by 10.514.725.2+10.5
- MedXPertQA MMQwen3.7 Plus Preview by 9.761.371+9.7
- HLE-DiamondQwen3.7 Plus Preview by 6.77.213.9+6.7
- GDP (Surge AI)Qwen3.7 Plus Preview by 6.2612.2+6.2
- SuperGPQAQwen3.7 Plus Preview by 5.765.771.4+5.7
- MATH-VisionQwen3.7 Plus Preview by 4.785.690.3+4.7
- MMLU-ProQwen3.7 Plus Preview by 3.385.288.5+3.3
- VITA-BenchQwen3.7 Plus Preview by 2.64345.6+2.6
- MMLU-Reduxtie93.794.5tie
- MMMLUtie88.489tie
Questions people ask
Which is better, Gemma 4 31B or Qwen3.7 Plus Preview?
Qwen3.7 Plus Preview wins all eight areas we test: coding, agents, reasoning, facts, images and charts, math, long documents and following instructions. Gemma 4 31B wins none, but costs 72% less.
Which is better for coding?
Qwen3.7 Plus Preview. It wins 6 of the 7 coding tests both models report; Gemma 4 31B wins none, and 1 is a tie.
Which is cheaper?
Gemma 4 31B costs $0.17 per million input tokens and $0.40 per million output tokens; Qwen3.7 Plus Preview costs $0.40 and $1.60. That makes Gemma 4 31B about 72% cheaper for the same work.
How do you compare the two?
We use the 46 benchmark tests both models have published scores on. The verdict counts the 23 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 23 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.