Model comparison
DeepSeek-V3 vs GLM-4.7
GLM-4.7 is the stronger model overall, scoring 42.0 to 39.5 on the Noometry Index. DeepSeek-V3 costs 2.5× less per token, which makes it the better buy when GLM-4.7's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
DeepSeekSummary
- They share 26 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and GLM-4.7 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GLM-4.7 leads 47.0 to 37.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 83.3% for GLM-4.7.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.60 / $2.20 for GLM-4.7.
- GLM-4.7 accepts more context: 205K tokens versus 164K.
Side by side
| DeepSeek-V3 | GLM-4.7 | |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 39.5 | 42.0 |
| Released | 2024-12-26 | 2025-12-22 |
| Weights | Open | Open |
| Context window | 164K | 205K |
| Max output | 164K | 131K |
| Input $ / M tokens | $0.24 | $0.60 |
| Output $ / M tokens | $0.90 | $2.20 |
| Results tracked | 60 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-4.7 leads
DeepSeek-V3: 42.3 (#106), GLM-4.7: 44.0 (#79)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| SciCode | 35.8% | 45.1% |
| LMArena Coding | 1368 | 1454 |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1435 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 399.48 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, GLM-4.7: 26.5 (#103)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| Terminal-Bench | — | 33.4% |
| METR Time Horizons | 49.6% | — |
| Vending-Bench 2 | — | 2,377 |
Reasoning GLM-4.7 leads
DeepSeek-V3: 20.5 (#236), GLM-4.7: 24.3 (#164)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| SimpleBench | 27.2% | 47.7% |
| CritPt | 0% | 1.7% |
| LMArena Hard Prompts | 1365 | 1443 |
| Epoch Capabilities Index | 135.94 | 143.51 |
| Kagi LLM Benchmark | 52.3% | — |
| Chess Puzzles | — | 6% |
| LiveBench Reasoning | 65.8% | — |
| DTBench | 64.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| LMCA | 15.5% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math GLM-4.7 leads
DeepSeek-V3: 32.1 (#219), GLM-4.7: 38.6 (#135)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 83.3% |
| LMArena Math | 1373 | 1423 |
| FrontierMath (Feb 2025 set) | 1.7% | 2.4% |
| ProofBench | — | 6% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge GLM-4.7 leads
DeepSeek-V3: 37.5 (#155), GLM-4.7: 47.0 (#80)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| GPQA Diamond | 67.6% | 83.3% |
| Vectara Hallucination Rate | 6.1% | 11.7% |
| LMArena Expert | 1351 | 1424 |
| SimpleQA Verified | — | 32.2% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multilingual GLM-4.7 leads
DeepSeek-V3: 48.5 (#143), GLM-4.7: 52.8 (#79)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| LMArena Non-English | 1358 | 1417 |
| LMArena Chinese | 1391 | 1495 |
| LMArena French | 1385 | 1432 |
| LMArena German | 1374 | 1424 |
| LMArena Japanese | 1333 | 1439 |
| LMArena Korean | 1319 | 1399 |
| LMArena Russian | 1373 | 1423 |
| LMArena Spanish | 1358 | 1434 |
Instruction Following GLM-4.7 leads
DeepSeek-V3: 72.8 (#130), GLM-4.7: 74.4 (#95)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| LMArena Instruction Following | 1345 | 1411 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context GLM-4.7 leads
DeepSeek-V3: 34.0 (#253), GLM-4.7: 42.8 (#116)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| LMArena Longer Query | 1352 | 1432 |
| Fiction.LiveBench | 50% | — |
| CL-bench | — | 15.9% |
| CL-bench Life | — | 10.9% |
Writing & Preference GLM-4.7 leads
DeepSeek-V3: 57.4 (#130), GLM-4.7: 60.9 (#93)
| Benchmark | DeepSeek-V3 | GLM-4.7 |
|---|---|---|
| LMArena Text | 1375 | 1435 |
| LMArena Creative Writing | 1364 | 1401 |
| EQ-Bench Creative Writing | 1472 | 1413 |
| LMArena Multi-Turn | 1389 | 1446 |
| Short-Story Creative Writing | 77% | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than GLM-4.7?
GLM-4.7 is the stronger model overall, scoring 42.0 to 39.5 on the Noometry Index. DeepSeek-V3 costs 2.5× less per token, which makes it the better buy when GLM-4.7's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3 or GLM-4.7?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; GLM-4.7 lists at $0.60 and $2.20.
Is DeepSeek-V3 or GLM-4.7 better for coding?
GLM-4.7 scores higher on coding benchmarks: 44.0 versus 42.3 in the Noometry coding category.
Which has the bigger context window?
GLM-4.7 does, with 205K tokens against 164K.
How many benchmarks do DeepSeek-V3 and GLM-4.7 share?
26 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and GLM-4.7 has 36.