GPT-5.2 vs Qwen3.6 35B A3B
Wins 6 of 8 areas
Coding · Agents · Reasoning · Math · Long documents · Following instructions
Wins 1 of 8 areas
Images and charts
GPT-5.2 is the stronger all-rounder.Qwen3.6 35B A3B is cheaper and better at images and charts.
Scores updated · 39 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.
- MathCompetition and research-level math40GPT-5.24 of 4 tests
- ReasoningHard problems that need careful thinking30GPT-5.23 of 3 tests
- CodingWriting and fixing software31GPT-5.23 of 5 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own31GPT-5.23 of 4 tests
- Long documentsFinding answers in very long texts10GPT-5.21 of 1 test
- Following instructionsDoing exactly what it is asked10GPT-5.21 of 1 test
- Images and chartsUnderstanding pictures, charts and video12Qwen3.6 35B A3B2 of 4 tests · 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.
Qwen3.6 35B A3B 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
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+47points ahead
- Real work tasks from 44 professionsGDPVal+28.5points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+25.5points ahead
Where Qwen3.6 35B A3B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+30.7points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+19.6points ahead
- Reads and reasons about text in imagesOCRBenchv2+15points ahead
Every test, side by side
All 39 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.2
- SWE-bench ProQwen3.6 35B A3B by 19.629.949.5+19.6
- SciCodeGPT-5.2 by 15.552.136.6+15.5
- Terminal-Bench HardGPT-5.2 by 12.24734.8+12.2
- SWE-bench VerifiedGPT-5.2 by 6.68073.4+6.6
- SWE-bench Multilingualtie66.767.2tie
AgentsGPT-5.2
- GDPValGPT-5.2 by 28.548.319.8+28.5
- BrowseCompGPT-5.2 by 17.165.848.7+17.1
- MCP AtlasQwen3.6 35B A3B by 2.260.662.8+2.2
- τ-Bench V3 · BankingGPT-5.2 by 1.811.19.3+1.8
ReasoningGPT-5.2
- Humanity's Last ExamGPT-5.2 by 15.537.722.2+15.5
- CritPtGPT-5.2 by 11.311.60.3+11.3
- GPQA DiamondGPT-5.2 by 6.290.384.1+6.2
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.6 35B A3B by 30.718.849.5+30.7
- AA-Omniscience · AccuracyGPT-5.2 by 25.544.318.8+25.5
Images and chartsQwen3.6 35B A3B
- OCRBenchv2Qwen3.6 35B A3B by 1550.565.5+15
- CharXiv (RQ)GPT-5.2 by 4.182.178+4.1
- MathVistaQwen3.6 35B A3B by 3.682.886.4+3.6
- MMMU-Protie74.675tie
MathGPT-5.2
- FrontierMath Tiers 1-3 (v2)GPT-5.2 by 4767.420.4+47
- HMMT Feb. 2026GPT-5.2 by 13.49783.6+13.4
- IMOAnswerBenchGPT-5.2 by 7.486.378.9+7.4
- AIME 2026GPT-5.2 by 5.698.392.7+5.6
Other results15 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 45.260.215+45.2
- AA-OmniscienceGPT-5.2 by 21.3-0.9-22.2+21.3
- ToolathlonGPT-5.2 by 19.446.326.9+19.4
- AA IntelligenceGPT-5.2 by 12.230.418.2+12.2
- τ²-Bench Telecom (AA run)Qwen3.6 35B A3B by 10.584.895.3+10.5
- HMMT 2025GPT-5.2 by 10.399.489.1+10.3
- HMMT Feb. 2025GPT-5.2 by 9.310090.7+9.3
- Artificial Analysis Coding IndexGPT-5.2 by 6.848.741.9+6.8
- HMMT Nov. 2025GPT-5.2 by 6.795.889.1+6.7
- GPQA (unspecified)GPT-5.2 by 6.492.486+6.4
- OmniDocBench 1.5Qwen3.6 35B A3B by 4.285.789.9+4.2
- SimpleVQAQwen3.6 35B A3B by 3.155.858.9+3.1
- Terminal-Bench 2.0GPT-5.2 by 2.55451.5+2.5
- VideoMMMUGPT-5.2 by 2.285.983.7+2.2
- MMLU-ProGPT-5.2 by 1.586.785.2+1.5
Questions people ask
Which is better, GPT-5.2 or Qwen3.6 35B A3B?
GPT-5.2 wins six of the eight areas we test: coding, agents, reasoning, math, long documents and following instructions. Qwen3.6 35B A3B wins images and charts, and costs 83% less. They are level on facts.
Which is better for coding?
GPT-5.2. It wins 3 of the 5 coding tests both models report; Qwen3.6 35B A3B wins 1, 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; Qwen3.6 35B A3B costs $0.38 and $2.25. That makes Qwen3.6 35B A3B about 83% cheaper for the same work.
How do you compare the two?
We use the 39 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 15 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.