Model comparison
DeepSeek-V3.1-Terminus vs GPT-4.5
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 43.1 to 37.2 on the Noometry Index.
Last verified . 10 shared benchmarks.
DeepSeekSummary
- They share 10 benchmarks with published results for both. DeepSeek-V3.1-Terminus scores higher in 5 categories and GPT-4.5 in 2 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V3.1-Terminus leads 26.4 to 13.9.
- DeepSeek-V3.1-Terminus has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.1-Terminus | GPT-4.5 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 43.1 | 37.2 |
| Released | 2025-09-22 | 2025-02-27 |
| Weights | Open | Proprietary |
| Context window | 164K | — |
| Max output | 147K | — |
| Input $ / M tokens | $0.27 | — |
| Output $ / M tokens | $1 | — |
| Results tracked | 16 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-V3.1-Terminus: 42.0 (#113), GPT-4.5: 42.2 (#109)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Coding | 1426 | 1396 |
| Aider Polyglot | — | 44.9% |
| SciCode | 40.6% | — |
| WeirdML | — | 39.4% |
| LiveBench Coding | — | 75.2% |
| ALE-Bench | 745.17 | — |
Agentic & Tool Use Not comparable
DeepSeek-V3.1-Terminus: —, GPT-4.5: 27.9 (#97)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| Cybench | — | 17.5% |
Reasoning DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 26.4 (#133), GPT-4.5: 13.9 (#330)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Hard Prompts | 1426 | 1403 |
| ARC-AGI-2 | — | 0.8% |
| SimpleBench | — | 34.5% |
| Kagi LLM Benchmark | 57.4% | — |
| ARC-AGI-1 | — | 10.3% |
| CritPt | 1.7% | — |
| EnigmaEval | — | 3.2% |
| LiveBench Reasoning | — | 71.1% |
| DTBench | 81.3% | — |
| LiveBench Data Analysis | — | 64.3% |
| LMCA | 28.6% | — |
| Epoch Capabilities Index | — | 136.74 |
| ForecastBench | — | 61.7 |
| LiveBench | — | 69% |
Math DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 38.5 (#137), GPT-4.5: 32.6 (#211)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Math | 1402 | 1412 |
| OTIS Mock AIME 2024-2025 | — | 37.8% |
| LiveBench Math | — | 69.3% |
| MATH Level 5 | — | 78.6% |
Knowledge Not comparable
DeepSeek-V3.1-Terminus: —, GPT-4.5: 32.5 (#211)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| GPQA Diamond | — | 68.7% |
| Humanity's Last Exam | — | 5.4% |
| Confabulations | — | 13.6% |
| LMArena Expert | — | 1394 |
Multimodal Not comparable
DeepSeek-V3.1-Terminus: —, GPT-4.5: 37.6 (#71)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Vision | — | 1195 |
| VPCT | — | 45% |
Multilingual Too close to call
DeepSeek-V3.1-Terminus: 52.1 (#92), GPT-4.5: 52.5 (#83)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1407 | 1413 |
| LMArena Russian | 1436 | 1419 |
| LMArena Chinese | — | 1421 |
| LMArena French | — | 1418 |
| LMArena German | — | 1457 |
| LMArena Japanese | — | 1416 |
| LMArena Korean | — | 1392 |
Instruction Following DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 74.0 (#106), GPT-4.5: 72.6 (#134)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1404 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
Long Context DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 43.4 (#97), GPT-4.5: 40.4 (#155)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Longer Query | 1421 | 1406 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 61.0 (#92), GPT-4.5: 56.9 (#134)
| Benchmark | DeepSeek-V3.1-Terminus | GPT-4.5 |
|---|---|---|
| LMArena Text | 1419 | 1417 |
| LMArena Creative Writing | 1403 | 1394 |
| LMArena Multi-Turn | 1411 | 1444 |
| Short-Story Creative Writing | — | 75.6% |
| EQ-Bench Creative Writing | — | 1258 |
| LiveBench Language | — | 61.5% |
Frequently asked questions
Is DeepSeek-V3.1-Terminus better than GPT-4.5?
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 43.1 to 37.2 on the Noometry Index.
Is DeepSeek-V3.1-Terminus or GPT-4.5 better for coding?
They score almost the same on coding (42.0 vs 42.2); test both on your own repository before choosing.
How many benchmarks do DeepSeek-V3.1-Terminus and GPT-4.5 share?
10 benchmarks have published results for both models. DeepSeek-V3.1-Terminus has 16 scored results on Noometry and GPT-4.5 has 42.