Model comparison
GLM-5.3 vs GPT-5.6 Terra
GPT-5.6 Terra is the stronger model overall, scoring 59.1 to 54.6 on the Noometry Index. GLM-5.3 costs 2.1× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Last verified . 39 shared benchmarks.
Summary
- They share 39 benchmarks with published results for both. GLM-5.3 scores higher in 5 categories and GPT-5.6 Terra in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Terra leads 81.5 to 62.1.
- The biggest single-benchmark swing is FrontierMath Tier 4: 29.3% for GLM-5.3 and 70.7% for GPT-5.6 Terra.
- GLM-5.3 is cheaper at $1.40 / $4.40 per million input/output tokens, against $2 / $12 for GPT-5.6 Terra.
- GPT-5.6 Terra 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-5.6 Terra | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 54.6 | 59.1 |
| Released | 2026-08-14 | 2026-07-09 |
| Weights | Open | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 131K | 128K |
| Input $ / M tokens | $1.40 | $2 |
| Output $ / M tokens | $4.40 | $12 |
| Results tracked | 41 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3 leads
GLM-5.3: 59.0 (#12), GPT-5.6 Terra: 57.1 (#16)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| DeepSWE | 69% | 69.6% |
| FrontierCode | 40.1% | 41.3% |
| LMArena WebDev | 1623 | 1520 |
| SciCode | 59% | 55% |
| WeirdML | 75.4% | 78.3% |
| LMArena Coding | 1499 | 1484 |
| ALE-Bench | 1,317 | 1,951 |
| CursorBench | — | 41.3% |
| FrontierSWE | 30.2% | — |
Agentic & Tool Use GPT-5.6 Terra leads
GLM-5.3: 36.3 (#34), GPT-5.6 Terra: 40.2 (#23)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| APEX-Agents | 56.6% | 58.2% |
| Vending-Bench 2 | 8,164 | 7,343 |
| BALROG | — | 53.2% |
| GDP.pdf | — | 24.7% |
Reasoning GPT-5.6 Terra leads
GLM-5.3: 45.2 (#42), GPT-5.6 Terra: 60.6 (#19)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| NYT Connections (extended) | 74.2% | 78.4% |
| CritPt | 19.1% | 30% |
| Chess Puzzles | 21% | 54% |
| LMArena Hard Prompts | 1491 | 1467 |
| Mystery Game Puzzles | 33% | 35% |
| DTBench | 87.7% | 93.3% |
| LMCA | 55.5% | 55% |
| Epoch Capabilities Index | 155.56 | 159.31 |
| ARC-AGI-2 | — | 83.9% |
| SimpleBench | — | 48.9% |
| Kagi LLM Benchmark | — | 51.3% |
| ARC-AGI-1 | — | 96.5% |
| Surface Evolver Bench | — | 83.8% |
| Bench to the Future 3 | 0.15 | — |
Math GPT-5.6 Terra leads
GLM-5.3: 62.1 (#27), GPT-5.6 Terra: 81.5 (#10)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| FrontierMath (Tiers 1-3) | 68.8% | 86% |
| FrontierMath Tier 4 | 29.3% | 70.7% |
| OTIS Mock AIME 2024-2025 | 91.1% | 99.7% |
| ProofBench | 49% | 74% |
| LMArena Math | 1493 | 1463 |
Knowledge GPT-5.6 Terra leads
GLM-5.3: 58.3 (#34), GPT-5.6 Terra: 61.3 (#27)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| GPQA Diamond | 90.9% | 93.3% |
| SimpleQA Verified | 41% | 43.2% |
| LMArena Expert | 1512 | 1492 |
Multimodal Not comparable
GLM-5.3: —, GPT-5.6 Terra: 47.5 (#10)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Vision | — | 1271 |
| Blueprint-Bench 2 | — | 30.8% |
| Furniture Assembly | — | 54.2% |
| LMArena Document | — | 1472 |
Multilingual GLM-5.3 leads
GLM-5.3: 55.9 (#27), GPT-5.6 Terra: 54.5 (#42)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Non-English | 1457 | 1439 |
| LMArena Chinese | 1523 | 1514 |
| LMArena French | 1499 | 1476 |
| LMArena German | 1505 | 1462 |
| LMArena Japanese | 1438 | 1446 |
| LMArena Korean | 1478 | 1420 |
| LMArena Russian | 1460 | 1450 |
| LMArena Spanish | 1444 | 1455 |
Instruction Following GLM-5.3 leads
GLM-5.3: 77.8 (#15), GPT-5.6 Terra: 76.6 (#38)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Instruction Following | 1481 | 1453 |
Long Context GLM-5.3 leads
GLM-5.3: 45.5 (#32), GPT-5.6 Terra: 44.4 (#64)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Longer Query | 1485 | 1452 |
Writing & Preference GLM-5.3 leads
GLM-5.3: 75.9 (#6), GPT-5.6 Terra: 70.4 (#21)
| Benchmark | GLM-5.3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Text | 1473 | 1446 |
| LMArena Creative Writing | 1459 | 1411 |
| EQ-Bench Creative Writing | 2075 | 1855 |
| LMArena Multi-Turn | 1473 | 1451 |
| EQ-Bench 4 | — | 1234 |
Frequently asked questions
Is GLM-5.3 better than GPT-5.6 Terra?
GPT-5.6 Terra is the stronger model overall, scoring 59.1 to 54.6 on the Noometry Index. GLM-5.3 costs 2.1× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Which is cheaper, GLM-5.3 or GPT-5.6 Terra?
GLM-5.3 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; GPT-5.6 Terra lists at $2 and $12.
Is GLM-5.3 or GPT-5.6 Terra better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.0 versus 57.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Terra does, with 1.05M tokens against 1M.
How many benchmarks do GLM-5.3 and GPT-5.6 Terra share?
39 benchmarks have published results for both models. GLM-5.3 has 41 scored results on Noometry and GPT-5.6 Terra has 52.