Model comparison
GLM-5.3 vs gpt-oss-120b
GLM-5.3 is the stronger model overall, scoring 54.8 to 36.3 on the Noometry Index. gpt-oss-120b costs 31× less per token, which makes it the better buy when GLM-5.3's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. GLM-5.3 scores higher in 9 categories and gpt-oss-120b in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GLM-5.3 leads 75.7 to 46.5.
- The biggest single-benchmark swing is APEX-Agents: 56.6% for GLM-5.3 and 4.4% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $1.40 / $4.40 for GLM-5.3.
- GLM-5.3 accepts more context: 1M tokens versus 131K.
Side by side
| GLM-5.3 | gpt-oss-120b | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 54.8 | 36.3 |
| Released | 2026-08-14 | 2025-08-05 |
| Weights | Open | Open |
| Context window | 1M | 131K |
| Max output | 131K | 41K |
| Input $ / M tokens | $1.40 | $0.037 |
| Output $ / M tokens | $4.40 | $0.17 |
| Results tracked | 42 | 48 |
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-oss-120b: 33.5 (#256)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| SciCode | 59% | 36% |
| WeirdML | 75.4% | 48.2% |
| LMArena Coding | 1496 | 1380 |
| ALE-Bench | 1,317 | 575.62 |
| DeepSWE | 69% | — |
| FrontierCode | 40.1% | — |
| SWE-bench Verified (bash only) | — | 26% |
| Aider Polyglot | — | 41.8% |
| CursorBench | 42.6% | — |
| LMArena WebDev | 1622 | — |
| FrontierSWE | 30.2% | — |
| AlgoTune | — | 1.41 |
Agentic & Tool Use GLM-5.3 leads
GLM-5.3: 36.4 (#38), gpt-oss-120b: 12.2 (#153)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| APEX-Agents | 56.6% | 4.4% |
| Vending-Bench 2 | 8,164 | -21.53 |
| Terminal-Bench | — | 18.7% |
| METR Time Horizons | — | 56.6% |
Reasoning GLM-5.3 leads
GLM-5.3: 46.1 (#46), gpt-oss-120b: 20.0 (#245)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| CritPt | 19.1% | 1.1% |
| Chess Puzzles | 21% | 20% |
| LMArena Hard Prompts | 1489 | 1364 |
| Mystery Game Puzzles | 33% | 2% |
| DTBench | 87.7% | 76.3% |
| LMCA | 55.5% | 22.1% |
| Epoch Capabilities Index | 155.61 | 139.93 |
| SimpleBench | — | 22.1% |
| Kagi LLM Benchmark | — | 58.6% |
| NYT Connections (extended) | 74.2% | — |
| Surface Evolver Bench | — | 25% |
| Bench to the Future 3 | 0.15 | — |
Math GLM-5.3 leads
GLM-5.3: 62.3 (#33), gpt-oss-120b: 52.5 (#50)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.1% | 88.9% |
| LMArena Math | 1489 | 1389 |
| FrontierMath (Tiers 1-3) | 68.8% | — |
| FrontierMath Tier 4 | 29.3% | — |
| ProofBench | 49% | — |
| Omni-MATH | — | 68.8% |
Knowledge GLM-5.3 leads
GLM-5.3: 58.3 (#37), gpt-oss-120b: 42.4 (#96)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | 90.9% | 75.8% |
| LMArena Expert | 1516 | 1356 |
| SimpleQA Verified | 41% | — |
| MMLU-Pro | — | 79.5% |
| Confabulations | — | 15.7% |
| Vectara Hallucination Rate | — | 14.2% |
| GPQA (HELM) | — | 68.4% |
Multilingual GLM-5.3 leads
GLM-5.3: 55.7 (#28), gpt-oss-120b: 48.0 (#147)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1457 | 1351 |
| LMArena Chinese | 1528 | 1385 |
| LMArena French | 1499 | 1369 |
| LMArena German | 1499 | 1353 |
| LMArena Japanese | 1453 | 1331 |
| LMArena Korean | 1472 | 1282 |
| LMArena Russian | 1463 | 1343 |
| LMArena Spanish | 1460 | 1389 |
Instruction Following GLM-5.3 leads
GLM-5.3: 77.5 (#23), gpt-oss-120b: 69.3 (#173)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| LMArena Instruction Following | 1477 | 1318 |
| IFEval | — | 83.6% |
Long Context GLM-5.3 leads
GLM-5.3: 45.4 (#41), gpt-oss-120b: 31.4 (#278)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| LMArena Longer Query | 1482 | 1319 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference GLM-5.3 leads
GLM-5.3: 75.7 (#6), gpt-oss-120b: 46.5 (#217)
| Benchmark | GLM-5.3 | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1471 | 1365 |
| LMArena Creative Writing | 1457 | 1275 |
| EQ-Bench Creative Writing | 2075 | 961 |
| LMArena Multi-Turn | 1472 | 1340 |
| Short-Story Creative Writing | — | 77.1% |
| WildBench | — | 84.5% |
Frequently asked questions
Is GLM-5.3 better than gpt-oss-120b?
GLM-5.3 is the stronger model overall, scoring 54.8 to 36.3 on the Noometry Index. gpt-oss-120b costs 31× 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-oss-120b?
gpt-oss-120b is cheaper. It lists at $0.037 per million input tokens and $0.17 per million output tokens; GLM-5.3 lists at $1.40 and $4.40.
Is GLM-5.3 or gpt-oss-120b better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.5 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
GLM-5.3 does, with 1M tokens against 131K.
How many benchmarks do GLM-5.3 and gpt-oss-120b share?
31 benchmarks have published results for both models. GLM-5.3 has 42 scored results on Noometry and gpt-oss-120b has 48.