Model comparison
DeepSeek-V3.2-Exp vs GLM-4.7
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 42.0 on the Noometry Index.
Last verified . 33 shared benchmarks.
DeepSeekSummary
- They share 33 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 7 categories and GLM-4.7 in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where DeepSeek-V3.2-Exp leads 32.7 to 26.5.
- The biggest single-benchmark swing is Chess Puzzles: 14% for DeepSeek-V3.2-Exp and 6% for GLM-4.7.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 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-V3.2-Exp | GLM-4.7 | |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 44.3 | 42.0 |
| Released | 2025-09-29 | 2025-12-22 |
| Weights | Open | Open |
| Context window | 164K | 205K |
| Max output | 66K | 131K |
| Input $ / M tokens | $0.26 | $0.60 |
| Output $ / M tokens | $0.38 | $2.20 |
| Results tracked | 49 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), GLM-4.7: 44.0 (#79)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| LMArena WebDev | 1362 | 1435 |
| SciCode | 38.9% | 45.1% |
| LMArena Coding | 1454 | 1454 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| SWE-bench Multilingual | 59% | — |
| WeirdML | 39.5% | — |
| ALE-Bench | — | 399.48 |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 32.7 (#59), GLM-4.7: 26.5 (#103)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| Terminal-Bench | 39.6% | 33.4% |
| Vending-Bench 2 | 1,034 | 2,377 |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
Reasoning GLM-4.7 leads
DeepSeek-V3.2-Exp: 22.1 (#208), GLM-4.7: 24.3 (#164)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| CritPt | 2.9% | 1.7% |
| Chess Puzzles | 14% | 6% |
| LMArena Hard Prompts | 1434 | 1443 |
| Epoch Capabilities Index | 146.27 | 143.51 |
| ARC-AGI-2 | 4% | — |
| SimpleBench | — | 47.7% |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| Thematic Generalization | 65% | — |
| DTBench | 87.7% | — |
| LMCA | 29.1% | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), GLM-4.7: 38.6 (#135)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 83.3% |
| ProofBench | 8% | 6% |
| LMArena Math | 1435 | 1423 |
| FrontierMath (Feb 2025 set) | 22.1% | 2.4% |
| FrontierMath Tier 4 (v1) | 2.1% | 0% |
| MathArena Final-Answer Competitions | 57.7% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), GLM-4.7: 47.0 (#80)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| GPQA Diamond | 83.4% | 83.3% |
| Vectara Hallucination Rate | 5.3% | 11.7% |
| LMArena Expert | 1436 | 1424 |
| SimpleQA Verified | — | 32.2% |
Multilingual Too close to call
DeepSeek-V3.2-Exp: 52.2 (#90), GLM-4.7: 52.8 (#79)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| LMArena Non-English | 1409 | 1417 |
| LMArena Chinese | 1461 | 1495 |
| LMArena French | 1433 | 1432 |
| LMArena German | 1440 | 1424 |
| LMArena Japanese | 1374 | 1439 |
| LMArena Korean | 1371 | 1399 |
| LMArena Russian | 1424 | 1423 |
| LMArena Spanish | 1440 | 1434 |
Instruction Following Too close to call
DeepSeek-V3.2-Exp: 74.5 (#93), GLM-4.7: 74.4 (#95)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| LMArena Instruction Following | 1413 | 1411 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), GLM-4.7: 42.8 (#116)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| CL-bench | 13.2% | 15.9% |
| CL-bench Life | 9.5% | 10.9% |
| LMArena Longer Query | 1428 | 1432 |
| Fiction.LiveBench | 83.3% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), GLM-4.7: 60.9 (#93)
| Benchmark | DeepSeek-V3.2-Exp | GLM-4.7 |
|---|---|---|
| LMArena Text | 1425 | 1435 |
| LMArena Creative Writing | 1403 | 1401 |
| EQ-Bench Creative Writing | 1515 | 1413 |
| LMArena Multi-Turn | 1427 | 1446 |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than GLM-4.7?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 42.0 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or GLM-4.7?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; GLM-4.7 lists at $0.60 and $2.20.
Is DeepSeek-V3.2-Exp or GLM-4.7 better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 44.0 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-V3.2-Exp and GLM-4.7 share?
33 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and GLM-4.7 has 36.