Model comparison
DeepSeek-V3.2-Exp vs GPT-4.5
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 37.2 on the Noometry Index.
Last verified . 25 shared benchmarks.
DeepSeekSummary
- They share 25 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 8 categories and GPT-4.5 in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 32.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 87.8% for DeepSeek-V3.2-Exp and 37.8% for GPT-4.5.
- DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.2-Exp | GPT-4.5 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 44.3 | 37.2 |
| Released | 2025-09-29 | 2025-02-27 |
| Weights | Open | Proprietary |
| Context window | 164K | — |
| Max output | 66K | — |
| Input $ / M tokens | $0.26 | — |
| Output $ / M tokens | $0.38 | — |
| Results tracked | 49 | 42 |
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), GPT-4.5: 42.2 (#109)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| Aider Polyglot | 74.2% | 44.9% |
| WeirdML | 39.5% | 39.4% |
| LMArena Coding | 1454 | 1396 |
| SWE-bench Verified (bash only) | 70% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| LiveBench Coding | — | 75.2% |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 32.7 (#59), GPT-4.5: 27.9 (#97)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Cybench | — | 17.5% |
| Vending-Bench 2 | 1,034 | — |
Reasoning DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 22.1 (#208), GPT-4.5: 13.9 (#330)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| ARC-AGI-2 | 4% | 0.8% |
| ARC-AGI-1 | 57% | 10.3% |
| LMArena Hard Prompts | 1434 | 1403 |
| Epoch Capabilities Index | 146.27 | 136.74 |
| SimpleBench | — | 34.5% |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| CritPt | 2.9% | — |
| Chess Puzzles | 14% | — |
| EnigmaEval | — | 3.2% |
| Thematic Generalization | 65% | — |
| LiveBench Reasoning | — | 71.1% |
| DTBench | 87.7% | — |
| LiveBench Data Analysis | — | 64.3% |
| LMCA | 29.1% | — |
| ForecastBench | — | 61.7 |
| LiveBench | — | 69% |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), GPT-4.5: 32.6 (#211)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 37.8% |
| LMArena Math | 1435 | 1412 |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| LiveBench Math | — | 69.3% |
| MATH Level 5 | — | 78.6% |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), GPT-4.5: 32.5 (#211)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| GPQA Diamond | 83.4% | 68.7% |
| LMArena Expert | 1436 | 1394 |
| Humanity's Last Exam | — | 5.4% |
| Confabulations | — | 13.6% |
| Vectara Hallucination Rate | 5.3% | — |
Multimodal Not comparable
DeepSeek-V3.2-Exp: —, GPT-4.5: 37.6 (#71)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| LMArena Vision | — | 1195 |
| VPCT | — | 45% |
Multilingual Too close to call
DeepSeek-V3.2-Exp: 52.2 (#90), GPT-4.5: 52.5 (#83)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1409 | 1413 |
| LMArena Chinese | 1461 | 1421 |
| LMArena French | 1433 | 1418 |
| LMArena German | 1440 | 1457 |
| LMArena Japanese | 1374 | 1416 |
| LMArena Korean | 1371 | 1392 |
| LMArena Russian | 1424 | 1419 |
| LMArena Spanish | 1440 | — |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), GPT-4.5: 72.6 (#134)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1413 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), GPT-4.5: 40.4 (#155)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| Fiction.LiveBench | 83.3% | 63.9% |
| LMArena Longer Query | 1428 | 1406 |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), GPT-4.5: 56.9 (#134)
| Benchmark | DeepSeek-V3.2-Exp | GPT-4.5 |
|---|---|---|
| LMArena Text | 1425 | 1417 |
| LMArena Creative Writing | 1403 | 1394 |
| EQ-Bench Creative Writing | 1515 | 1258 |
| LMArena Multi-Turn | 1427 | 1444 |
| Short-Story Creative Writing | — | 75.6% |
| LiveBench Language | — | 61.5% |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than GPT-4.5?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 37.2 on the Noometry Index.
Is DeepSeek-V3.2-Exp or GPT-4.5 better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 42.2 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.2-Exp and GPT-4.5 share?
25 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and GPT-4.5 has 42.