GPT-5.2 vs Qwen3.7 Plus Preview
Wins 3 of 8 areas
Reasoning · Math · Long documents
Wins 3 of 8 areas
Agents · Images and charts · Following instructions
The two are evenly matched.GPT-5.2 is better at reasoning and math; Qwen3.7 Plus Preview at images and charts and agents.
Scores updated · 38 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.
- Images and chartsUnderstanding pictures, charts and video05Qwen3.7 Plus Preview5 of 5 tests
- AgentsCarrying out multi-step tasks on its own12Qwen3.7 Plus Preview2 of 4 tests · 1 tie
- Following instructionsDoing exactly what it is asked01Qwen3.7 Plus Preview1 of 1 test
- ReasoningHard problems that need careful thinking20GPT-5.22 of 3 tests · 1 tie
- MathCompetition and research-level math20GPT-5.22 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts10GPT-5.21 of 1 test
- CodingWriting and fixing software22Even2 each · 1 tie
- FactsGetting facts right instead of making them up11Even1 each
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.
Qwen3.7 Plus Preview costs 87% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.7 Plus Preview pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+53.5points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+27.7points ahead
- Reads and reasons about text in imagesOCRBenchv2+16.6points ahead
Where GPT-5.2 pulls ahead
- Real work tasks from 44 professionsGDPVal+34.9points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+33points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+21.8points ahead
Every test, side by side
All 38 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingEven
- SWE-bench ProQwen3.7 Plus Preview by 27.729.957.6+27.7
- SWE-bench MultilingualQwen3.7 Plus Preview by 9.166.775.8+9.1
- SciCodeGPT-5.2 by 652.146.1+6
- SWE-bench VerifiedGPT-5.2 by 2.38077.7+2.3
- Terminal-Bench Hardtie4747tie
AgentsQwen3.7 Plus Preview
- GDPValGPT-5.2 by 34.948.313.4+34.9
- MCP AtlasQwen3.7 Plus Preview by 12.660.673.2+12.6
- τ-Bench V3 · BankingQwen3.7 Plus Preview by 6.411.117.5+6.4
- AA ApexAgentstie2322.4tie
ReasoningGPT-5.2
- CritPtGPT-5.2 by 2.511.69.1+2.5
- Humanity's Last ExamGPT-5.2 by 2.137.735.6+2.1
- GPQA Diamondtie90.390tie
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.7 Plus Preview by 53.518.872.3+53.5
- AA-Omniscience · AccuracyGPT-5.2 by 21.844.322.5+21.8
Images and chartsQwen3.7 Plus Preview
- OCRBenchv2Qwen3.7 Plus Preview by 16.650.567.1+16.6
- MMMU-ProQwen3.7 Plus Preview by 5.974.680.5+5.9
- CharXiv (RQ)Qwen3.7 Plus Preview by 3.882.185.9+3.8
- LMArena · VisionQwen3.7 Plus Preview by 32 rating points12451277+32 rating
- Video-MMEQwen3.7 Plus Preview by 28688+2
MathGPT-5.2
- FrontierMath Tiers 1-3 (v2)GPT-5.2 by 3367.434.4+33
- HMMT Feb. 2026GPT-5.2 by 4.19792.9+4.1
- IMOAnswerBenchtie86.386tie
Following instructionsQwen3.7 Plus Preview
- IFBenchQwen3.7 Plus Preview by 2.675.478+2.6
Other results14 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 Agentic IndexGPT-5.2 by 40.560.219.7+40.5
- WorldVQAQwen3.7 Plus Preview by 33.12861.1+33.1
- SimpleVQAQwen3.7 Plus Preview by 25.955.881.7+25.9
- Terminal-Bench 2.0Qwen3.7 Plus Preview by 16.35470.3+16.3
- τ²-Bench Telecom (AA run)Qwen3.7 Plus Preview by 8.284.893+8.2
- MATH-VisionQwen3.7 Plus Preview by 7.38390.3+7.3
- ScreenSpot-Pro (No tools)GPT-5.2 by 7.386.379+7.3
- Artificial Analysis Coding IndexQwen3.7 Plus Preview by 7.248.755.9+7.2
- OmniDocBench 1.5Qwen3.7 Plus Preview by 5.785.791.4+5.7
- AA IntelligenceGPT-5.2 by 5.230.425.2+5.2
- AA-OmniscienceQwen3.7 Plus Preview by 2-0.91.1+2
- MMLU-ProQwen3.7 Plus Preview by 1.886.788.5+1.8
- MMMLUtie89.689tie
- VideoMMMUtie85.985.4tie
Questions people ask
Which is better, GPT-5.2 or Qwen3.7 Plus Preview?
GPT-5.2 and Qwen3.7 Plus Preview each win three of the eight areas we test. GPT-5.2 wins reasoning, math and long documents; Qwen3.7 Plus Preview wins agents, images and charts and following instructions. Qwen3.7 Plus Preview costs 87% less. They are level on coding and facts.
Which is better for coding?
Neither. They win 2 coding tests each of the 5 both models report, and 1 is a tie.
Which is cheaper?
GPT-5.2 costs $1.75 per million input tokens and $14.00 per million output tokens; Qwen3.7 Plus Preview costs $0.40 and $1.60. That makes Qwen3.7 Plus Preview about 87% cheaper for the same work.
How do you compare the two?
We use the 38 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 14 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.