GLM-5.3-Flash vs GPT-5.6 Luna
Wins 3 of 8 areas
Coding · Agents · Images and charts
Wins 5 of 8 areas
Reasoning · Facts · Math · Long documents · Following instructions
GPT-5.6 Luna is the stronger all-rounder.GLM-5.3-Flash is cheaper and better at 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.
- MathCompetition and research-level math03GPT-5.6 Luna3 of 3 tests
- ReasoningHard problems that need careful thinking12GPT-5.6 Luna2 of 5 tests · 2 ties
- FactsGetting facts right instead of making them up12GPT-5.6 Luna2 of 3 tests
- Long documentsFinding answers in very long texts01GPT-5.6 Luna1 of 1 test
- Following instructionsDoing exactly what it is asked01GPT-5.6 Luna1 of 1 test
- AgentsCarrying out multi-step tasks on its own42GLM-5.3-Flash4 of 6 tests
- Images and chartsUnderstanding pictures, charts and video20GLM-5.3-Flash2 of 2 tests
- CodingWriting and fixing software32GLM-5.3-Flash3 of 5 tests
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-Flash costs 54% 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-5.6 Luna pulls ahead
- Completes tasks by operating a computer desktopOSWorld-Verified+57.8points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+41.4points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+26.3points ahead
Where GLM-5.3-Flash pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+65points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+21.2points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+16.1points 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.
CodingGLM-5.3-Flash
- LMArena · WebDevGLM-5.3-Flash by 90 rating points16091519+90 rating
- LiveBench · Agentic CodingGLM-5.3-Flash by 8.456.848.4+8.4
- LiveBench · CodingGPT-5.6 Luna by 3.97982.9+3.9
- Terminal-Bench 2.1GLM-5.3-Flash by 3.484.380.9+3.4
- SciCodeGPT-5.6 Luna by 251.653.6+2
AgentsGLM-5.3-Flash
- OSWorld-VerifiedGPT-5.6 Luna by 57.814.872.6+57.8
- Terminal-Bench 4.0GLM-5.3-Flash by 21.232.811.6+21.2
- τ-Bench V3 · BankingGLM-5.3-Flash by 16.147.231.1+16.1
- AA IT-Bench SREGLM-5.3-Flash by 10.951.240.3+10.9
- GDPValGLM-5.3-Flash by 9.257.248+9.2
- BrowseCompGPT-5.6 Luna by 7.37683.3+7.3
ReasoningGPT-5.6 Luna
- LiveBench · ReasoningGPT-5.6 Luna by 877.685.6+8
- ARC-AGI-2GLM-5.3-Flash by 6.265.859.6+6.2
- CritPtGPT-5.6 Luna by 5.215.420.6+5.2
- Humanity's Last Examtie39.939.5tie
- GPQA Diamondtie91.291.1tie
FactsGPT-5.6 Luna
- AA-Omniscience · Non-hallucinationGLM-5.3-Flash by 6572.47.4+65
- AA-Omniscience · AccuracyGPT-5.6 Luna by 15.227.542.7+15.2
- SimpleQA VerifiedGPT-5.6 Luna by 93241+9
Images and chartsGLM-5.3-Flash
- CharXiv (RQ)GLM-5.3-Flash by 6.789.482.7+6.7
- LMArena · VisionGLM-5.3-Flash by 39 rating points12951256+39 rating
MathGPT-5.6 Luna
- FrontierMath Tier 4GPT-5.6 Luna by 41.417.158.5+41.4
- FrontierMath Tiers 1-3 (v2)GPT-5.6 Luna by 26.355.882.1+26.3
- LiveBench · MathematicsGPT-5.6 Luna by 681.287.2+6
Following instructionsGPT-5.6 Luna
- LiveBench · Instruction FollowingGPT-5.6 Luna by 7.352.860.1+7.3
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.
- ToolathlonGLM-5.3-Flash by 2578.453.4+25
- Agents' Last ExamGPT-5.6 Luna by 2426.350.3+24
- AA-OmniscienceGLM-5.3-Flash by 17.87.5-10.3+17.8
- MV-BenchGLM-5.3-Flash by 9.677.868.2+9.6
- AA Agentic IndexGLM-5.3-Flash by 8.551.242.7+8.5
- GDP (Surge AI)GPT-5.6 Luna by 7.315.422.7+7.3
- livebench_languageGLM-5.3-Flash by 4.777.372.6+4.7
- AA IntelligenceGLM-5.3-Flash by 4.541.837.3+4.5
- DeepSWE 1.1GPT-5.6 Luna by 3.663.467+3.6
- livebench_data_analysisGPT-5.6 Luna by 1.676.478+1.6
- ARC-AGI-1tie9190.7tie
- Artificial Analysis Coding Indextie71.571.4tie
Questions people ask
Which is better, GLM-5.3-Flash or GPT-5.6 Luna?
GPT-5.6 Luna wins five of the eight areas we test: reasoning, facts, math, long documents and following instructions. GLM-5.3-Flash wins coding, agents and images and charts, and costs 54% less.
Which is better for coding?
GLM-5.3-Flash. It wins 3 of the 5 coding tests both models report; GPT-5.6 Luna wins 2.
Which is cheaper?
GLM-5.3-Flash costs $0.15 per million input tokens and $0.50 per million output tokens; GPT-5.6 Luna costs $0.20 and $1.20. That makes GLM-5.3-Flash about 54% 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 26 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.