Model comparison
GLM-5 vs GPT-5.6 Luna
GPT-5.6 Luna is the stronger model overall, scoring 54.6 to 46.1 on the Noometry Index.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. GLM-5 scores higher in 2 categories and GPT-5.6 Luna in 7 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Luna leads 77.7 to 46.4.
- The biggest single-benchmark swing is ARC-AGI-2: 4.9% for GLM-5 and 59.5% for GPT-5.6 Luna.
- GPT-5.6 Luna is cheaper at $0.20 / $1.20 per million input/output tokens, against $1 / $3.20 for GLM-5.
- GPT-5.6 Luna accepts more context: 1.05M tokens versus 205K.
- GLM-5 has downloadable open weights; the other is API-only.
Side by side
| GLM-5 | GPT-5.6 Luna | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 46.1 | 54.6 |
| Released | 2026-02-11 | 2026-07-09 |
| Weights | Open | Proprietary |
| Context window | 205K | 1.05M |
| Max output | 131K | 128K |
| Input $ / M tokens | $1 | $0.20 |
| Output $ / M tokens | $3.20 | $1.20 |
| Results tracked | 45 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Luna leads
GLM-5: 49.0 (#52), GPT-5.6 Luna: 54.5 (#28)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| LMArena WebDev | 1434 | 1519 |
| WeirdML | 48.2% | 60.9% |
| LMArena Coding | 1461 | 1466 |
| ALE-Bench | 765.62 | 1,667 |
| SWE-bench Verified | 72.1% | — |
| DeepSWE | — | 67.2% |
| FrontierCode | — | 39.8% |
| SWE-bench Verified (bash only) | 72.8% | — |
| CursorBench | — | 35.9% |
| SWE-bench Multilingual | 69.7% | — |
| SciCode | — | 53.6% |
Agentic & Tool Use GPT-5.6 Luna leads
GLM-5: 31.1 (#71), GPT-5.6 Luna: 34.4 (#45)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| Vending-Bench 2 | 4,432 | 4,095 |
| Terminal-Bench | 52.4% | — |
| APEX-Agents | — | 43% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 9.8% | — |
| τ²-bench Retail | 73.7% | — |
| τ²-bench Telecom | 86.8% | — |
| BALROG | — | 45.6% |
| GDP.pdf | — | 22.7% |
Reasoning GPT-5.6 Luna leads
GLM-5: 27.6 (#116), GPT-5.6 Luna: 47.6 (#43)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| ARC-AGI-2 | 4.9% | 59.5% |
| SimpleBench | 53.2% | 46.8% |
| Kagi LLM Benchmark | 75% | 49.1% |
| NYT Connections (extended) | 74.8% | 69.4% |
| ARC-AGI-1 | 44.7% | 88% |
| Chess Puzzles | 10% | 40% |
| LMArena Hard Prompts | 1452 | 1451 |
| Epoch Capabilities Index | 145.83 | 156.39 |
| CritPt | — | 20.6% |
| Mystery Game Puzzles | — | 21% |
| DTBench | — | 89.1% |
| LMCA | — | 48.5% |
| Surface Evolver Bench | — | 61.9% |
| ForecastBench | 61 | — |
Math GPT-5.6 Luna leads
GLM-5: 46.4 (#71), GPT-5.6 Luna: 77.7 (#14)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 80% | 98.3% |
| LMArena Math | 1440 | 1458 |
| FrontierMath (Tiers 1-3) | — | 82.1% |
| FrontierMath Tier 4 | — | 61% |
| MathArena Final-Answer Competitions | 65.7% | — |
| ProofBench | — | 60% |
| FrontierMath (Feb 2025 set) | 16.4% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5.6 Luna leads
GLM-5: 52.3 (#64), GPT-5.6 Luna: 58.5 (#34)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| GPQA Diamond | 87.8% | 91.6% |
| LMArena Expert | 1454 | 1478 |
| SimpleQA Verified | — | 41% |
| Vectara Hallucination Rate | 10.1% | — |
Multimodal Not comparable
GLM-5: —, GPT-5.6 Luna: 42.7 (#28)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Vision | — | 1258 |
| Blueprint-Bench 2 | — | 22.6% |
| Furniture Assembly | — | 42.5% |
| LMArena Document | — | 1457 |
Multilingual Too close to call
GLM-5: 53.7 (#58), GPT-5.6 Luna: 52.8 (#78)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Non-English | 1430 | 1417 |
| LMArena Chinese | 1511 | 1470 |
| LMArena French | 1455 | 1456 |
| LMArena German | 1445 | 1454 |
| LMArena Japanese | 1416 | 1411 |
| LMArena Korean | 1423 | 1415 |
| LMArena Russian | 1436 | 1428 |
| LMArena Spanish | 1454 | 1448 |
Instruction Following Too close to call
GLM-5: 75.2 (#67), GPT-5.6 Luna: 75.6 (#57)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Instruction Following | 1428 | 1437 |
Long Context Too close to call
GLM-5: 44.7 (#60), GPT-5.6 Luna: 43.9 (#82)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Longer Query | 1446 | 1436 |
| CL-bench | 18.7% | — |
Writing & Preference GPT-5.6 Luna leads
GLM-5: 66.0 (#38), GPT-5.6 Luna: 68.0 (#29)
| Benchmark | GLM-5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Text | 1446 | 1431 |
| LMArena Creative Writing | 1439 | 1396 |
| EQ-Bench Creative Writing | 1601 | 1829 |
| LMArena Multi-Turn | 1456 | 1434 |
| EQ-Bench 4 | — | 1156 |
Frequently asked questions
Is GLM-5 better than GPT-5.6 Luna?
GPT-5.6 Luna is the stronger model overall, scoring 54.6 to 46.1 on the Noometry Index.
Which is cheaper, GLM-5 or GPT-5.6 Luna?
GPT-5.6 Luna is cheaper. It lists at $0.20 per million input tokens and $1.20 per million output tokens; GLM-5 lists at $1 and $3.20.
Is GLM-5 or GPT-5.6 Luna better for coding?
GPT-5.6 Luna scores higher on coding benchmarks: 54.5 versus 49.0 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Luna does, with 1.05M tokens against 205K.
How many benchmarks do GLM-5 and GPT-5.6 Luna share?
31 benchmarks have published results for both models. GLM-5 has 45 scored results on Noometry and GPT-5.6 Luna has 52.