GPT-5.2 vs Qwen3.5 27B
Wins 5 of 8 areas
Coding · Agents · Reasoning · Facts · Math
Wins 1 of 8 areas
Images and charts
GPT-5.2 is the stronger all-rounder.Qwen3.5 27B is cheaper and better at images and charts.
Scores updated · 47 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.
- ReasoningHard problems that need careful thinking30GPT-5.23 of 3 tests
- MathCompetition and research-level math30GPT-5.23 of 3 tests
- CodingWriting and fixing software42GPT-5.24 of 6 tests
- FactsGetting facts right instead of making them up20GPT-5.22 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own21GPT-5.22 of 3 tests
- Images and chartsUnderstanding pictures, charts and video12Qwen3.5 27B2 of 5 tests · 2 ties
- Long documentsFinding answers in very long texts11Even1 each
- Following instructionsDoing exactly what it is asked00Even0 each · 1 tie
Following instructions rests on a single test.
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.
Qwen3.5 27B costs 83% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-5.2 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+23.6points ahead
- Harvard-MIT high-school math contest problemsHMMT Feb. 2026+15.9points ahead
- Real work tasks from 44 professionsGDPVal+15.3points ahead
Where Qwen3.5 27B pulls ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+21.3points ahead
- Reads text in imagesOCRBench+8.7points ahead
- Multi-step tasks using many real software toolsMCP Atlas+7.8points ahead
Every test, side by side
All 47 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.2
- SWE-bench ProQwen3.5 27B by 21.329.951.2+21.3
- Terminal-Bench HardGPT-5.2 by 14.44732.6+14.4
- SciCodeGPT-5.2 by 12.652.139.5+12.6
- SWE-bench VerifiedGPT-5.2 by 7.68072.4+7.6
- LMArena · WebDevGPT-5.2 by 59 rating points14161357+59 rating
- SWE-bench MultilingualQwen3.5 27B by 2.666.769.3+2.6
AgentsGPT-5.2
- GDPValGPT-5.2 by 15.348.333+15.3
- MCP AtlasQwen3.5 27B by 7.860.668.4+7.8
- BrowseCompGPT-5.2 by 4.865.861+4.8
ReasoningGPT-5.2
- Humanity's Last ExamGPT-5.2 by 13.837.723.9+13.8
- CritPtGPT-5.2 by 10.711.60.9+10.7
- GPQA DiamondGPT-5.2 by 4.590.385.8+4.5
FactsGPT-5.2
- AA-Omniscience · AccuracyGPT-5.2 by 23.644.320.7+23.6
- Vectara HHEM hallucination ratelower is betterGPT-5.2 by 1.310.812.1+1.3
- AA-Omniscience · Non-hallucinationtie18.818.5tie
Images and chartsQwen3.5 27B
- OCRBenchQwen3.5 27B by 8.780.789.4+8.7
- MathVistaQwen3.5 27B by 582.887.8+5
- CharXiv (RQ)GPT-5.2 by 2.682.179.5+2.6
- LMArena · Visiontie12461240tie
- MMMU-Protie74.675tie
MathGPT-5.2
- HMMT Feb. 2026GPT-5.2 by 15.99781.1+15.9
- AIME 2026GPT-5.2 by 6.698.391.7+6.6
- IMOAnswerBenchGPT-5.2 by 6.486.379.9+6.4
Long documentsEven
- LongBench v2Qwen3.5 27B by 6.154.560.6+6.1
- AA-LCRGPT-5.2 by 582.777.7+5
Other results21 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-5.2 by 43.1-0.9-44+43.1
- ScreenSpot-Pro (No tools)GPT-5.2 by 1686.370.3+16
- Artificial Analysis Coding IndexGPT-5.2 by 13.848.734.9+13.8
- Terminal-Bench 2.0GPT-5.2 by 12.45441.6+12.4
- HMMT 2025GPT-5.2 by 9.699.489.8+9.6
- vectara_avg_summary_lengthGPT-5.2 by 91.9 rating points186.394.4+91.9 rating
- τ²-Bench Telecom (AA run)Qwen3.5 27B by 9.184.893.9+9.1
- HMMT Feb. 2025GPT-5.2 by 810092+8
- AA IntelligenceGPT-5.2 by 7.530.422.9+7.5
- MMVUGPT-5.2 by 7.580.873.3+7.5
- HMMT Nov. 2025GPT-5.2 by 695.889.8+6
- AA Agentic IndexGPT-5.2 by 5.660.254.6+5.6
- MMMLUGPT-5.2 by 3.789.685.9+3.7
- VideoMMMUGPT-5.2 by 3.685.982.3+3.6
- OmniDocBench 1.5Qwen3.5 27B by 3.285.788.9+3.2
- MATH-VisionQwen3.5 27B by 38386+3
- vectara_factual_consistencyGPT-5.2 by 1.389.287.9+1.3
- ZeroBenchQwen3.5 27B by 1910+1
- MMLU-Protie86.786.1tie
- SimpleVQAtie55.856tie
- vectara_answer_ratetie10099.8tie
Questions people ask
Which is better, GPT-5.2 or Qwen3.5 27B?
GPT-5.2 wins five of the eight areas we test: coding, agents, reasoning, facts and math. Qwen3.5 27B wins images and charts, and costs 83% less. They are level on long documents and following instructions.
Which is better for coding?
GPT-5.2. It wins 4 of the 6 coding tests both models report; Qwen3.5 27B wins 2.
Which is cheaper?
GPT-5.2 costs $1.75 per million input tokens and $14.00 per million output tokens; Qwen3.5 27B costs $0.30 and $2.40. That makes Qwen3.5 27B about 83% cheaper for the same work.
How do you compare the two?
We use the 47 benchmark tests both models have published scores on. The verdict counts the 26 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 21 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.