Model comparison
GLM-5 vs GLM-5.3
GLM-5.3 is the stronger model overall, scoring 54.8 to 46.1 on the Noometry Index.
Last verified . 27 shared benchmarks.
Summary
- They share 27 benchmarks with published results for both. GLM-5 scores higher in 0 categories and GLM-5.3 in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-5.3 leads 46.1 to 27.6.
- The biggest single-benchmark swing is WeirdML: 48.2% for GLM-5 and 75.4% for GLM-5.3.
- GLM-5 is cheaper at $1 / $3.20 per million input/output tokens, against $1.40 / $4.40 for GLM-5.3.
- GLM-5.3 accepts more context: 1M tokens versus 205K.
Side by side
| GLM-5 | GLM-5.3 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Z.ai (Zhipu) |
| Noometry Index | 46.1 | 54.8 |
| Released | 2026-02-11 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 205K | 1M |
| Max output | 131K | 131K |
| Input $ / M tokens | $1 | $1.40 |
| Output $ / M tokens | $3.20 | $4.40 |
| Results tracked | 45 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3 leads
GLM-5: 49.0 (#52), GLM-5.3: 59.5 (#14)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| LMArena WebDev | 1434 | 1622 |
| WeirdML | 48.2% | 75.4% |
| LMArena Coding | 1461 | 1496 |
| ALE-Bench | 765.62 | 1,317 |
| SWE-bench Verified | 72.1% | — |
| DeepSWE | — | 69% |
| FrontierCode | — | 40.1% |
| SWE-bench Verified (bash only) | 72.8% | — |
| CursorBench | — | 42.6% |
| SWE-bench Multilingual | 69.7% | — |
| FrontierSWE | — | 30.2% |
| SciCode | — | 59% |
Agentic & Tool Use GLM-5.3 leads
GLM-5: 31.1 (#71), GLM-5.3: 36.4 (#38)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| Vending-Bench 2 | 4,432 | 8,164 |
| Terminal-Bench | 52.4% | — |
| APEX-Agents | — | 56.6% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 9.8% | — |
| τ²-bench Retail | 73.7% | — |
| τ²-bench Telecom | 86.8% | — |
Reasoning GLM-5.3 leads
GLM-5: 27.6 (#116), GLM-5.3: 46.1 (#46)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| NYT Connections (extended) | 74.8% | 74.2% |
| Chess Puzzles | 10% | 21% |
| LMArena Hard Prompts | 1452 | 1489 |
| Epoch Capabilities Index | 145.83 | 155.61 |
| ARC-AGI-2 | 4.9% | — |
| SimpleBench | 53.2% | — |
| Kagi LLM Benchmark | 75% | — |
| ARC-AGI-1 | 44.7% | — |
| CritPt | — | 19.1% |
| Mystery Game Puzzles | — | 33% |
| DTBench | — | 87.7% |
| LMCA | — | 55.5% |
| Bench to the Future 3 | — | 0.15 |
| ForecastBench | 61 | — |
Math GLM-5.3 leads
GLM-5: 46.4 (#71), GLM-5.3: 62.3 (#33)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 80% | 91.1% |
| LMArena Math | 1440 | 1489 |
| FrontierMath (Tiers 1-3) | — | 68.8% |
| FrontierMath Tier 4 | — | 29.3% |
| MathArena Final-Answer Competitions | 65.7% | — |
| ProofBench | — | 49% |
| FrontierMath (Feb 2025 set) | 16.4% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GLM-5.3 leads
GLM-5: 52.3 (#64), GLM-5.3: 58.3 (#37)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| GPQA Diamond | 87.8% | 90.9% |
| LMArena Expert | 1454 | 1516 |
| SimpleQA Verified | — | 41% |
| Vectara Hallucination Rate | 10.1% | — |
Multilingual GLM-5.3 leads
GLM-5: 53.7 (#58), GLM-5.3: 55.7 (#28)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| LMArena Non-English | 1430 | 1457 |
| LMArena Chinese | 1511 | 1528 |
| LMArena French | 1455 | 1499 |
| LMArena German | 1445 | 1499 |
| LMArena Japanese | 1416 | 1453 |
| LMArena Korean | 1423 | 1472 |
| LMArena Russian | 1436 | 1463 |
| LMArena Spanish | 1454 | 1460 |
Instruction Following GLM-5.3 leads
GLM-5: 75.2 (#67), GLM-5.3: 77.5 (#23)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| LMArena Instruction Following | 1428 | 1477 |
Long Context Too close to call
GLM-5: 44.7 (#60), GLM-5.3: 45.4 (#41)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| LMArena Longer Query | 1446 | 1482 |
| CL-bench | 18.7% | — |
Writing & Preference GLM-5.3 leads
GLM-5: 66.0 (#38), GLM-5.3: 75.7 (#6)
| Benchmark | GLM-5 | GLM-5.3 |
|---|---|---|
| LMArena Text | 1446 | 1471 |
| LMArena Creative Writing | 1439 | 1457 |
| EQ-Bench Creative Writing | 1601 | 2075 |
| LMArena Multi-Turn | 1456 | 1472 |
Frequently asked questions
Is GLM-5 better than GLM-5.3?
GLM-5.3 is the stronger model overall, scoring 54.8 to 46.1 on the Noometry Index.
Which is cheaper, GLM-5 or GLM-5.3?
GLM-5 is cheaper. It lists at $1 per million input tokens and $3.20 per million output tokens; GLM-5.3 lists at $1.40 and $4.40.
Is GLM-5 or GLM-5.3 better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.5 versus 49.0 in the Noometry coding category.
Which has the bigger context window?
GLM-5.3 does, with 1M tokens against 205K.
How many benchmarks do GLM-5 and GLM-5.3 share?
27 benchmarks have published results for both models. GLM-5 has 45 scored results on Noometry and GLM-5.3 has 42.