Model comparison
GLM-5.3 vs GPT-6 Luna
GLM-5.3 is the stronger model overall, scoring 54.8 to 53.3 on the Noometry Index. GPT-6 Luna costs 11× less per token, which makes it the better buy when GLM-5.3's lead doesn't matter for your workload.
Last verified . 36 shared benchmarks.
Summary
- They share 36 benchmarks with published results for both. GLM-5.3 scores higher in 7 categories and GPT-6 Luna in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GLM-5.3 leads 75.7 to 58.3.
- The biggest single-benchmark swing is FrontierMath Tier 4: 29.3% for GLM-5.3 and 56.1% for GPT-6 Luna.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $1.40 / $4.40 for GLM-5.3.
- GPT-6 Luna accepts more context: 1.05M tokens versus 1M.
- GLM-5.3 has downloadable open weights; the other is API-only.
Side by side
| GLM-5.3 | GPT-6 Luna | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 54.8 | 53.3 |
| Released | 2026-08-14 | 2026-09-22 |
| Weights | Open | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 131K | 128K |
| Input $ / M tokens | $1.40 | $0.10 |
| Output $ / M tokens | $4.40 | $0.50 |
| Results tracked | 42 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3 leads
GLM-5.3: 59.5 (#14), GPT-6 Luna: 55.5 (#25)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| DeepSWE | 69% | 66.6% |
| FrontierCode | 40.1% | 42.4% |
| LMArena WebDev | 1622 | 1581 |
| SciCode | 59% | 54.6% |
| LMArena Coding | 1496 | 1439 |
| ALE-Bench | 1,317 | 1,577 |
| CursorBench | 42.6% | — |
| FrontierSWE | 30.2% | — |
| WeirdML | 75.4% | — |
Agentic & Tool Use GLM-5.3 leads
GLM-5.3: 36.4 (#38), GPT-6 Luna: 33.3 (#54)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| APEX-Agents | 56.6% | 44.3% |
| GDP.pdf | — | 23% |
| Vending-Bench 2 | 8,164 | — |
Reasoning GPT-6 Luna leads
GLM-5.3: 46.1 (#46), GPT-6 Luna: 48.2 (#41)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| NYT Connections (extended) | 74.2% | 68.7% |
| CritPt | 19.1% | 19.4% |
| Chess Puzzles | 21% | 31% |
| LMArena Hard Prompts | 1489 | 1411 |
| Mystery Game Puzzles | 33% | 7% |
| DTBench | 87.7% | 90.1% |
| LMCA | 55.5% | 44.5% |
| Epoch Capabilities Index | 155.61 | 156.28 |
| ARC-AGI-2 | — | 59.3% |
| ARC-AGI-1 | — | 86.7% |
| Bench to the Future 3 | 0.15 | — |
Math GPT-6 Luna leads
GLM-5.3: 62.3 (#33), GPT-6 Luna: 76.1 (#15)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| FrontierMath (Tiers 1-3) | 68.8% | 78.9% |
| FrontierMath Tier 4 | 29.3% | 56.1% |
| OTIS Mock AIME 2024-2025 | 91.1% | 98.9% |
| ProofBench | 49% | 64% |
| LMArena Math | 1489 | 1416 |
Knowledge GLM-5.3 leads
GLM-5.3: 58.3 (#37), GPT-6 Luna: 57.0 (#41)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| GPQA Diamond | 90.9% | 90.5% |
| SimpleQA Verified | 41% | 41.4% |
| LMArena Expert | 1516 | 1444 |
Multimodal Not comparable
GLM-5.3: —, GPT-6 Luna: 42.4 (#30)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| LMArena Vision | — | 1217 |
| Blueprint-Bench 2 | — | 31.2% |
| Furniture Assembly | — | 44.2% |
Multilingual GLM-5.3 leads
GLM-5.3: 55.7 (#28), GPT-6 Luna: 50.5 (#117)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| LMArena Non-English | 1457 | 1386 |
| LMArena Chinese | 1528 | 1433 |
| LMArena French | 1499 | 1420 |
| LMArena German | 1499 | 1369 |
| LMArena Japanese | 1453 | 1369 |
| LMArena Korean | 1472 | 1360 |
| LMArena Russian | 1463 | 1394 |
| LMArena Spanish | 1460 | 1393 |
Instruction Following GLM-5.3 leads
GLM-5.3: 77.5 (#23), GPT-6 Luna: 74.3 (#99)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| LMArena Instruction Following | 1477 | 1409 |
Long Context GLM-5.3 leads
GLM-5.3: 45.4 (#41), GPT-6 Luna: 43.0 (#111)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| LMArena Longer Query | 1482 | 1409 |
Writing & Preference GLM-5.3 leads
GLM-5.3: 75.7 (#6), GPT-6 Luna: 58.3 (#119)
| Benchmark | GLM-5.3 | GPT-6 Luna |
|---|---|---|
| LMArena Text | 1471 | 1391 |
| LMArena Creative Writing | 1457 | 1363 |
| LMArena Multi-Turn | 1472 | 1396 |
| EQ-Bench Creative Writing | 2075 | — |
Frequently asked questions
Is GLM-5.3 better than GPT-6 Luna?
GLM-5.3 is the stronger model overall, scoring 54.8 to 53.3 on the Noometry Index. GPT-6 Luna costs 11× less per token, which makes it the better buy when GLM-5.3's lead doesn't matter for your workload.
Which is cheaper, GLM-5.3 or GPT-6 Luna?
GPT-6 Luna is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; GLM-5.3 lists at $1.40 and $4.40.
Is GLM-5.3 or GPT-6 Luna better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.5 versus 55.5 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Luna does, with 1.05M tokens against 1M.
How many benchmarks do GLM-5.3 and GPT-6 Luna share?
36 benchmarks have published results for both models. GLM-5.3 has 42 scored results on Noometry and GPT-6 Luna has 42.