Model comparison
DeepSeek-V2.5 (Sep 2024) vs GPT-4.5
DeepSeek-V2.5 (Sep 2024) and GPT-4.5 score almost the same on the Noometry Index (37.0 vs 36.9), so choose on price, context window or the category you care about most.
Last verified . 17 shared benchmarks.
DeepSeekSummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 3 categories and GPT-4.5 in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-4.5 leads 41.8 to 31.3.
- The biggest single-benchmark swing is Aider Polyglot: 17.8% for DeepSeek-V2.5 (Sep 2024) and 44.9% for GPT-4.5.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | GPT-4.5 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 37.0 | 36.9 |
| Released | 2024-09-06 | 2025-02-27 |
| Weights | Open | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4.5 leads
DeepSeek-V2.5 (Sep 2024): 31.3 (#274), GPT-4.5: 41.8 (#102)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| Aider Polyglot | 17.8% | 44.9% |
| LMArena Coding | 1310 | 1397 |
| WeirdML | — | 39.4% |
| BigCodeBench Instruct | 48.6% | — |
| LiveBench Coding | — | 75.2% |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, GPT-4.5: 27.6 (#92)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| Cybench | — | 17.5% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 24.5 (#145), GPT-4.5: 14.1 (#320)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1404 |
| ARC-AGI-2 | — | 0.8% |
| SimpleBench | — | 34.5% |
| ARC-AGI-1 | — | 10.3% |
| EnigmaEval | — | 3.2% |
| LiveBench Reasoning | — | 71.1% |
| LiveBench Data Analysis | — | 64.3% |
| Epoch Capabilities Index | — | 136.74 |
| ForecastBench | — | 61.7 |
| LiveBench | — | 69% |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 33.3 (#177), GPT-4.5: 31.0 (#200)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Math | 1289 | 1412 |
| OTIS Mock AIME 2024-2025 | — | 37.8% |
| LiveBench Math | — | 69.3% |
| MATH Level 5 | — | 78.6% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.2 (#190), GPT-4.5: 32.4 (#205)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Expert | 1266 | 1394 |
| GPQA Diamond | — | 68.7% |
| Humanity's Last Exam | — | 5.4% |
| Confabulations | — | 13.6% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, GPT-4.5: 37.0 (#68)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Vision | — | 1195 |
| VPCT | — | 45% |
Multilingual GPT-4.5 leads
DeepSeek-V2.5 (Sep 2024): 42.7 (#188), GPT-4.5: 52.7 (#78)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1273 | 1413 |
| LMArena Chinese | 1317 | 1421 |
| LMArena French | 1288 | 1417 |
| LMArena German | 1258 | 1457 |
| LMArena Japanese | 1228 | 1416 |
| LMArena Korean | 1209 | 1392 |
| LMArena Russian | 1289 | 1419 |
| LMArena Spanish | 1248 | — |
Instruction Following GPT-4.5 leads
DeepSeek-V2.5 (Sep 2024): 67.7 (#189), GPT-4.5: 72.8 (#129)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1280 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
Long Context Too close to call
DeepSeek-V2.5 (Sep 2024): 39.5 (#169), GPT-4.5: 40.4 (#150)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Longer Query | 1301 | 1406 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference GPT-4.5 leads
DeepSeek-V2.5 (Sep 2024): 50.0 (#182), GPT-4.5: 57.1 (#129)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.5 |
|---|---|---|
| LMArena Text | 1294 | 1417 |
| LMArena Creative Writing | 1285 | 1394 |
| LMArena Multi-Turn | 1297 | 1444 |
| Short-Story Creative Writing | — | 75.6% |
| EQ-Bench Creative Writing | — | 1258 |
| LiveBench Language | — | 61.5% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than GPT-4.5?
DeepSeek-V2.5 (Sep 2024) and GPT-4.5 score almost the same on the Noometry Index (37.0 vs 36.9), so choose on price, context window or the category you care about most.
Is DeepSeek-V2.5 (Sep 2024) or GPT-4.5 better for coding?
GPT-4.5 scores higher on coding benchmarks: 41.8 versus 31.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and GPT-4.5 share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and GPT-4.5 has 42.