Model comparison
Gemini 3 Pro vs GLM-5.3-Flash
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 51.8 on the Noometry Index.
Last verified . 28 shared benchmarks.
GoogleSummary
- They share 28 benchmarks with published results for both. Gemini 3 Pro scores higher in 6 categories and GLM-5.3-Flash in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in multimodal, where Gemini 3 Pro leads 57.6 to 42.8.
- The biggest single-benchmark swing is ARC-AGI-2: 31.1% for Gemini 3 Pro and 65.8% for GLM-5.3-Flash.
- GLM-5.3-Flash has downloadable open weights; the other is API-only.
Side by side
| Gemini 3 Pro | GLM-5.3-Flash | |
|---|---|---|
| Provider | Z.ai (Zhipu) | |
| Noometry Index | 54.8 | 51.8 |
| Released | 2025-11-18 | 2026-08-20 |
| Weights | Proprietary | Open |
| Context window | — | 1M |
| Max output | — | 131K |
| Input $ / M tokens | — | $0.15 |
| Output $ / M tokens | — | $0.50 |
| Results tracked | 67 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3-Flash leads
Gemini 3 Pro: 51.6 (#39), GLM-5.3-Flash: 53.1 (#31)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| LMArena WebDev | 1440 | 1609 |
| LMArena Coding | 1481 | 1508 |
| ALE-Bench | 1,177 | 303.55 |
| SWE-bench Verified | 72.9% | — |
| DeepSWE | — | 63.4% |
| FrontierCode | — | 31.8% |
| SWE-bench Verified (bash only) | 74.2% | — |
| CursorBench | — | 36.8% |
| SWE-bench Multilingual | 68.7% | — |
| FrontierSWE | — | 18.1% |
| SciCode | — | 51.6% |
| GSO | 18.6% | — |
| WeirdML | 69.9% | — |
| AlgoTune | 1.83 | — |
Agentic & Tool Use Gemini 3 Pro leads
Gemini 3 Pro: 40.6 (#23), GLM-5.3-Flash: 34.2 (#47)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| Terminal-Bench | 69.4% | — |
| APEX-Agents | — | 52.8% |
| Berkeley Function Calling Leaderboard | 72.5% | — |
| GDPval | 40.3% | — |
| Remote Labor Index | 1.3% | — |
| τ²-bench Airline | 80.5% | — |
| τ²-bench Banking | 18% | — |
| τ²-bench Retail | 75.9% | — |
| τ²-bench Telecom | 91% | — |
| DeepResearch Bench | 46.3% | — |
| BALROG | 58.1% | — |
| GDP.pdf | — | 14% |
| LMArena Search | 1207 | — |
| METR Time Horizons | 71% | — |
| Vending-Bench 2 | 5,478 | — |
Reasoning Gemini 3 Pro leads
Gemini 3 Pro: 52.5 (#31), GLM-5.3-Flash: 48.0 (#42)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| ARC-AGI-2 | 31.1% | 65.8% |
| ARC-AGI-1 | 75% | 91% |
| CritPt | 6.9% | 15.4% |
| Chess Puzzles | 31% | 14% |
| LMArena Hard Prompts | 1480 | 1491 |
| Epoch Capabilities Index | 152.92 | 151.88 |
| SimpleBench | 76.4% | — |
| Kagi LLM Benchmark | 80.1% | — |
| NYT Connections (extended) | 94.4% | — |
| EnigmaEval | 18.2% | — |
| Mystery Game Puzzles | — | 8% |
| Surface Evolver Bench | — | 52.5% |
| Bench to the Future 3 | — | 0.15 |
| ForecastBench | 61.2 | — |
Math GLM-5.3-Flash leads
Gemini 3 Pro: 49.9 (#59), GLM-5.3-Flash: 53.3 (#47)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.4% | 93.9% |
| ProofBench | 20% | 21% |
| LMArena Math | 1476 | 1500 |
| FrontierMath (Tiers 1-3) | — | 55.8% |
| FrontierMath Tier 4 | — | 17.1% |
| MathArena Final-Answer Competitions | 67% | — |
| Omni-MATH | 55.5% | — |
| FrontierMath (Feb 2025 set) | 37.6% | — |
| FrontierMath Tier 4 (v1) | 18.8% | — |
Knowledge Gemini 3 Pro leads
Gemini 3 Pro: 64.4 (#16), GLM-5.3-Flash: 58.4 (#36)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| GPQA Diamond | 92.6% | 90.2% |
| LMArena Expert | 1475 | 1513 |
| Humanity's Last Exam | 37.5% | — |
| MMLU-Pro | 90.3% | — |
| Vectara Hallucination Rate | 13.6% | — |
| GPQA (HELM) | 80.3% | — |
Multimodal Gemini 3 Pro leads
Gemini 3 Pro: 57.6 (#2), GLM-5.3-Flash: 42.8 (#27)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| LMArena Vision | 1305 | 1296 |
| GeoBench | 84% | — |
| VPCT | 91% | — |
| LMArena Document | 1434 | — |
Multilingual Too close to call
Gemini 3 Pro: 56.9 (#16), GLM-5.3-Flash: 56.0 (#25)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| LMArena Non-English | 1474 | 1462 |
| LMArena Chinese | 1523 | 1527 |
| LMArena French | 1492 | 1496 |
| LMArena German | 1515 | 1470 |
| LMArena Japanese | 1510 | 1429 |
| LMArena Korean | 1448 | 1446 |
| LMArena Russian | 1493 | 1469 |
| LMArena Spanish | 1470 | 1471 |
Instruction Following GLM-5.3-Flash leads
Gemini 3 Pro: 76.3 (#45), GLM-5.3-Flash: 77.5 (#20)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| LMArena Instruction Following | 1458 | 1478 |
| IFEval | 87.7% | — |
Long Context GLM-5.3-Flash leads
Gemini 3 Pro: 44.0 (#79), GLM-5.3-Flash: 45.4 (#39)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| LMArena Longer Query | 1471 | 1482 |
| CL-bench | 15.8% | — |
Writing & Preference Gemini 3 Pro leads
Gemini 3 Pro: 66.4 (#35), GLM-5.3-Flash: 65.3 (#50)
| Benchmark | Gemini 3 Pro | GLM-5.3-Flash |
|---|---|---|
| LMArena Text | 1479 | 1471 |
| LMArena Creative Writing | 1482 | 1442 |
| LMArena Multi-Turn | 1484 | 1467 |
| EQ-Bench Creative Writing | 1525 | — |
| WildBench | 85.9% | — |
Frequently asked questions
Is Gemini 3 Pro better than GLM-5.3-Flash?
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 51.8 on the Noometry Index.
Is Gemini 3 Pro or GLM-5.3-Flash better for coding?
GLM-5.3-Flash scores higher on coding benchmarks: 53.1 versus 51.6 in the Noometry coding category.
How many benchmarks do Gemini 3 Pro and GLM-5.3-Flash share?
28 benchmarks have published results for both models. Gemini 3 Pro has 67 scored results on Noometry and GLM-5.3-Flash has 40.