GLM-5.2 vs GPT-5.2
Wins 3 of 7 areas
Coding · Agents · Reasoning
Wins 4 of 7 areas
Facts · Math · Long documents · Following instructions
GPT-5.2 wins more areas, narrowly.GLM-5.2 is cheaper and better at coding.
Scores updated · 40 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. 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 math14GPT-5.24 of 6 tests · 1 tie
- FactsGetting facts right instead of making them up12GPT-5.22 of 3 tests
- Following instructionsDoing exactly what it is asked01GPT-5.21 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts01GPT-5.21 of 1 test
- CodingWriting and fixing software50GLM-5.25 of 6 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own31GLM-5.23 of 4 tests
- ReasoningHard problems that need careful thinking32GLM-5.23 of 6 tests · 1 tie
Long documents 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.
GLM-5.2 costs 63% 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
- Abstract visual puzzles that people can solveARC-AGI-2+30.1points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+20points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+8.2points ahead
Where GLM-5.2 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+54.9points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+32.2points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+23.5points 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.
CodingGLM-5.2
- SWE-bench ProGLM-5.2 by 32.262.129.9+32.2
- LMArena · WebDevGLM-5.2 by 187 rating points16031416+187 rating
- Terminal-Bench HardGLM-5.2 by 3.850.847+3.8
- LiveBench · CodingGLM-5.2 by 3.679.776.1+3.6
- LiveBench · Agentic CodingGLM-5.2 by 1.551.850.3+1.5
- SciCodetie51.252.1tie
AgentsGLM-5.2
- τ-Bench V3 · BankingGLM-5.2 by 23.534.611.1+23.5
- MCP AtlasGLM-5.2 by 16.276.860.6+16.2
- AA ApexAgentsGLM-5.2 by 10.733.723+10.7
- GDPValGPT-5.2 by 4.643.748.3+4.6
ReasoningGLM-5.2
- ARC-AGI-2GPT-5.2 by 30.122.852.9+30.1
- SimpleBenchGLM-5.2 by 1358.845.8+13
- CritPtGLM-5.2 by 9.320.911.6+9.3
- LiveBench · ReasoningGPT-5.2 by 4.678.683.2+4.6
- Humanity's Last ExamGLM-5.2 by 3.441.137.7+3.4
- GPQA Diamondtie89.590.3tie
FactsGPT-5.2
- AA-Omniscience · Non-hallucinationGLM-5.2 by 54.973.718.8+54.9
- AA-Omniscience · AccuracyGPT-5.2 by 2024.344.3+20
- SimpleQA VerifiedGPT-5.2 by 2.934.237.1+2.9
MathGPT-5.2
- FrontierMath Tiers 1-3 (v2)GPT-5.2 by 8.259.267.4+8.2
- IMOAnswerBenchGLM-5.2 by 4.79186.3+4.7
- HMMT Feb. 2026GPT-5.2 by 4.592.597+4.5
- LiveBench · MathematicsGPT-5.2 by 3.489.893.2+3.4
- FrontierMath Tier 4GPT-5.2 by 2.429.331.7+2.4
- AIME 2026tie99.298.3tie
Following instructionsGPT-5.2
- IFBenchGPT-5.2 by 2.173.375.4+2.1
- LiveBench · Instruction Followingtie62.361.8tie
Other results12 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 20.839.460.2+20.8
- Artificial Analysis Coding IndexGLM-5.2 by 20.168.848.7+20.1
- τ²-Bench Telecom (AA run)GLM-5.2 by 14.399.184.8+14.3
- ARC-AGI-1GPT-5.2 by 9.27786.2+9.2
- HLE (with tools)GLM-5.2 by 9.254.745.5+9.2
- AA-OmniscienceGLM-5.2 by 5.34.4-0.9+5.3
- HMMT 2025GPT-5.2 by 594.499.4+5
- livebench_data_analysisGPT-5.2 by 4.573.778.2+4.5
- livebench_languageGPT-5.2 by 3.676.279.8+3.6
- AA IntelligenceGLM-5.2 by 3.333.730.4+3.3
- ToolathlonGLM-5.2 by 1.948.246.3+1.9
- HMMT Nov. 2025GPT-5.2 by 1.494.495.8+1.4
Questions people ask
Which is better, GLM-5.2 or GPT-5.2?
GPT-5.2 wins four of the seven areas where both have results: facts, math, long documents and following instructions. GLM-5.2 wins coding, agents and reasoning, and costs 63% less.
Which is better for coding?
GLM-5.2. It wins 5 of the 6 coding tests both models report; GPT-5.2 wins none, and 1 is a tie.
Which is cheaper?
GLM-5.2 costs $1.40 per million input tokens and $4.40 per million output tokens; GPT-5.2 costs $1.75 and $14.00. That makes GLM-5.2 about 63% 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 28 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 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.