GLM-5.3 vs Qwen3.8 Max Preview
Wins 1 of 7 areas
Agents
Wins 4 of 7 areas
Coding · Reasoning · Math · Following instructions
Qwen3.8 Max Preview wins more areas, narrowly.GLM-5.3 is cheaper and better at agents.
Both rank among the ten best models we track in agents.
Scores updated · 43 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 math03Qwen3.8 Max Preview3 of 3 tests
- CodingWriting and fixing software23Qwen3.8 Max Preview3 of 5 tests
- ReasoningHard problems that need careful thinking12Qwen3.8 Max Preview2 of 4 tests · 1 tie
- Following instructionsDoing exactly what it is asked01Qwen3.8 Max Preview1 of 1 test
- AgentsCarrying out multi-step tasks on its own30GLM-5.33 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up11Even1 each · 1 tie
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
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.
GLM-5.3 costs 27% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.8 Max Preview pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+17points ahead
- Short factual questions, answered correctlySimpleQA Verified+6.3points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+5.9points ahead
Where GLM-5.3 pulls ahead
- Code for real scientific research problemsSciCode+6.9points ahead
- Programming problems, refreshed regularlyLiveBench · Coding+6.1points ahead
- Finds the root cause of IT system incidentsAA IT-Bench SRE+5.8points ahead
Every test, side by side
All 43 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.8 Max Preview
- SciCodeGLM-5.3 by 6.95952.1+6.9
- LiveBench · CodingGLM-5.3 by 6.17972.9+6.1
- Terminal-Bench 2.1Qwen3.8 Max Preview by 4.983.988.8+4.9
- LMArena · WebDevQwen3.8 Max Preview by 48 rating points16221670+48 rating
- LiveBench · Agentic CodingQwen3.8 Max Preview by 3.860.964.7+3.8
AgentsGLM-5.3
- AA IT-Bench SREGLM-5.3 by 5.846.140.3+5.8
- Terminal-Bench 4.0GLM-5.3 by 341.938.9+3
- τ-Bench V3 · BankingGLM-5.3 by 2.550.347.8+2.5
- GDPValtie57.558.3tie
ReasoningQwen3.8 Max Preview
- LiveBench · ReasoningQwen3.8 Max Preview by 2.485.888.2+2.4
- CritPtGLM-5.3 by 1.419.117.7+1.4
- GPQA DiamondQwen3.8 Max Preview by 1.191.792.8+1.1
- Humanity's Last Examtie42.343.1tie
FactsEven
- SimpleQA VerifiedQwen3.8 Max Preview by 6.34147.3+6.3
- AA-Omniscience · AccuracyGLM-5.3 by 2.233.931.7+2.2
- AA-Omniscience · Non-hallucinationtie70.471.2tie
MathQwen3.8 Max Preview
- FrontierMath Tier 4Qwen3.8 Max Preview by 1729.346.3+17
- FrontierMath Tiers 1-3 (v2)Qwen3.8 Max Preview by 5.968.874.7+5.9
- LiveBench · MathematicsQwen3.8 Max Preview by 3.487.991.3+3.4
Following instructionsQwen3.8 Max Preview
- LiveBench · Instruction FollowingQwen3.8 Max Preview by 4.869.374.1+4.8
Other results22 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.
- Agents' Last ExamQwen3.8 Max Preview by 23.928.552.4+23.9
- GDP (Surge AI)Qwen3.8 Max Preview by 11.611.222.8+11.6
- DeepSWEGLM-5.3 by 10.366.956.6+10.3
- DeepSWE 1.1GLM-5.3 by 10.366.956.6+10.3
- ProgramBench (Almost Solved)GLM-5.3 by 8.51910.5+8.5
- livebench_data_analysisQwen3.8 Max Preview by 8.270.278.4+8.2
- HLE (with tools)GLM-5.3 by 6.362.556.2+6.3
- HLE-DiamondQwen3.8 Max Preview by 6.117.924+6.1
- CyberGymGLM-5.3 by 684.578.5+6
- HealthBenchQwen3.8 Max Preview by 5.754.560.2+5.7
- CritPt v2 HardGLM-5.3 by 433.529.5+4
- AA Agentic IndexGLM-5.3 by 3.853.449.6+3.8
- Artificial Analysis Coding IndexGLM-5.3 by 374.871.8+3
- GDPval-AA v2GLM-5.3 by 30 rating points17691739+30 rating
- Terminal-Bench-Science 0.1Qwen3.8 Max Preview by 2.49.511.9+2.4
- AA-OmniscienceGLM-5.3 by 2.314.312+2.3
- NL2RepoGLM-5.3 by 2.15855.9+2.1
- Agents' Last Exam (ALE-CLI)GLM-5.3 by 1.528.527+1.5
- AA Intelligencetie44.845.4tie
- Toolathlontie7372.5tie
- Toolathlon Verifiedtie7372.5tie
- livebench_languagetie79.979.7tie
Questions people ask
Which is better, GLM-5.3 or Qwen3.8 Max Preview?
Qwen3.8 Max Preview wins four of the seven areas where both have results: coding, reasoning, math and following instructions. GLM-5.3 wins agents, and costs 27% less. They are level on facts and long documents.
Which is better for coding?
Qwen3.8 Max Preview. It wins 3 of the 5 coding tests both models report; GLM-5.3 wins 2.
Which is cheaper?
GLM-5.3 costs $1.40 per million input tokens and $4.40 per million output tokens; Qwen3.8 Max Preview costs $2.00 and $6.00. That makes GLM-5.3 about 27% cheaper for the same work.
How do you compare the two?
We use the 43 benchmark tests both models have published scores on. The verdict counts the 21 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 22 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.