Model comparison
DeepSeek-V3 vs GLM-5.3-Flash
GLM-5.3-Flash is the stronger model overall, scoring 49.6 to 39.2 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeekSummary
- They share 22 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and GLM-5.3-Flash in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-5.3-Flash leads 53.1 to 30.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 93.9% for GLM-5.3-Flash.
- GLM-5.3-Flash is cheaper at $0.15 / $0.50 per million input/output tokens, against $0.24 / $0.90 for DeepSeek-V3.
- GLM-5.3-Flash accepts more context: 1M tokens versus 164K.
Side by side
| DeepSeek-V3 | GLM-5.3-Flash | |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 39.2 | 49.6 |
| Released | 2024-12-26 | 2026-08-20 |
| Weights | Open | Open |
| Context window | 164K | 1M |
| Max output | 164K | 131K |
| Input $ / M tokens | $0.24 | $0.15 |
| Output $ / M tokens | $0.90 | $0.50 |
| Results tracked | 60 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3-Flash leads
DeepSeek-V3: 41.9 (#99), GLM-5.3-Flash: 53.7 (#26)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| SciCode | 35.8% | 51.6% |
| LMArena Coding | 1368 | 1506 |
| DeepSWE | — | 63.4% |
| FrontierCode | — | 31.8% |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1614 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 303.55 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, GLM-5.3-Flash: 34.3 (#41)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| APEX-Agents | — | 52.8% |
| GDP.pdf | — | 14% |
| METR Time Horizons | 49.6% | — |
Reasoning GLM-5.3-Flash leads
DeepSeek-V3: 20.0 (#226), GLM-5.3-Flash: 31.7 (#69)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| CritPt | 0% | 15.4% |
| LMArena Hard Prompts | 1365 | 1490 |
| Epoch Capabilities Index | 135.95 | 151.9 |
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| Chess Puzzles | — | 14% |
| LiveBench Reasoning | 65.8% | — |
| Mystery Game Puzzles | — | 8% |
| DTBench | 64.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| LMCA | 15.5% | — |
| Surface Evolver Bench | — | 52.5% |
| Bench to the Future 3 | — | 0.15 |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math GLM-5.3-Flash leads
DeepSeek-V3: 30.4 (#208), GLM-5.3-Flash: 53.1 (#38)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 93.9% |
| LMArena Math | 1373 | 1499 |
| FrontierMath (Tiers 1-3) | — | 55.8% |
| FrontierMath Tier 4 | — | 17.1% |
| ProofBench | — | 21% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge GLM-5.3-Flash leads
DeepSeek-V3: 37.5 (#147), GLM-5.3-Flash: 58.4 (#33)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| GPQA Diamond | 67.6% | 90.2% |
| LMArena Expert | 1352 | 1512 |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multimodal Not comparable
DeepSeek-V3: —, GLM-5.3-Flash: 42.6 (#25)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| LMArena Vision | — | 1304 |
Multilingual GLM-5.3-Flash leads
DeepSeek-V3: 48.8 (#138), GLM-5.3-Flash: 56.2 (#24)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| LMArena Non-English | 1358 | 1462 |
| LMArena Chinese | 1390 | 1529 |
| LMArena French | 1386 | 1495 |
| LMArena German | 1374 | 1457 |
| LMArena Japanese | 1333 | 1420 |
| LMArena Korean | 1321 | 1443 |
| LMArena Russian | 1372 | 1472 |
| LMArena Spanish | 1358 | 1471 |
Instruction Following GLM-5.3-Flash leads
DeepSeek-V3: 73.0 (#125), GLM-5.3-Flash: 77.8 (#17)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| LMArena Instruction Following | 1345 | 1480 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context GLM-5.3-Flash leads
DeepSeek-V3: 34.0 (#248), GLM-5.3-Flash: 45.3 (#41)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| LMArena Longer Query | 1353 | 1479 |
| Fiction.LiveBench | 50% | — |
Writing & Preference GLM-5.3-Flash leads
DeepSeek-V3: 57.6 (#124), GLM-5.3-Flash: 65.4 (#49)
| Benchmark | DeepSeek-V3 | GLM-5.3-Flash |
|---|---|---|
| LMArena Text | 1375 | 1471 |
| LMArena Creative Writing | 1364 | 1440 |
| LMArena Multi-Turn | 1389 | 1471 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than GLM-5.3-Flash?
GLM-5.3-Flash is the stronger model overall, scoring 49.6 to 39.2 on the Noometry Index.
Which is cheaper, DeepSeek-V3 or GLM-5.3-Flash?
GLM-5.3-Flash is cheaper. It lists at $0.15 per million input tokens and $0.50 per million output tokens; DeepSeek-V3 lists at $0.24 and $0.90.
Is DeepSeek-V3 or GLM-5.3-Flash better for coding?
GLM-5.3-Flash scores higher on coding benchmarks: 53.7 versus 41.9 in the Noometry coding category.
Which has the bigger context window?
GLM-5.3-Flash does, with 1M tokens against 164K.
How many benchmarks do DeepSeek-V3 and GLM-5.3-Flash share?
22 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and GLM-5.3-Flash has 36.