Model comparison
Claude 3.5 Sonnet vs DeepSeek-V2.5 (Sep 2024)
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 34.6 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeekSummary
- They share 22 benchmarks with published results for both. Claude 3.5 Sonnet scores higher in 5 categories and DeepSeek-V2.5 (Sep 2024) in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 35.9 to 19.2.
- The biggest single-benchmark swing is Aider Polyglot: 51.6% for Claude 3.5 Sonnet and 17.8% for DeepSeek-V2.5 (Sep 2024).
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 34.6 | 37.6 |
| Released | 2024-06-20 | 2024-09-06 |
| Weights | Proprietary | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 60 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude 3.5 Sonnet leads
Claude 3.5 Sonnet: 39.0 (#165), DeepSeek-V2.5 (Sep 2024): 31.7 (#281)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| Aider Polyglot | 51.6% | 17.8% |
| BigCodeBench Instruct | 46.8% | 48.6% |
| LMArena Coding | 1342 | 1309 |
| BigCodeBench Complete | 58.6% | 53.2% |
| HumanEval+ | 81.7% | 83.5% |
| MBPP+ | 74.3% | 74.1% |
| GSO | 4.6% | — |
| WeirdML | 40% | — |
| LiveBench Coding | 67.1% | — |
| CadEval | 48% | — |
Agentic & Tool Use Not comparable
Claude 3.5 Sonnet: 32.3 (#67), DeepSeek-V2.5 (Sep 2024): —
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| TheAgentCompany | 24% | — |
| Cybench | 17.5% | — |
| BALROG | 32.6% | — |
| METR Time Horizons | 45.2% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
Claude 3.5 Sonnet: 23.1 (#183), DeepSeek-V2.5 (Sep 2024): 25.6 (#145)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Hard Prompts | 1305 | 1289 |
| SimpleBench | 41.4% | — |
| EnigmaEval | 0.9% | — |
| LiveBench Reasoning | 56.7% | — |
| DTBench | 67.8% | — |
| LiveBench Data Analysis | 55% | — |
| Epoch Capabilities Index | 133.55 | — |
| ForecastBench | 60.7 | — |
| LiveBench | 59% | — |
Math DeepSeek-V2.5 (Sep 2024) leads
Claude 3.5 Sonnet: 19.2 (#288), DeepSeek-V2.5 (Sep 2024): 35.9 (#177)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Math | 1307 | 1288 |
| OTIS Mock AIME 2024-2025 | 8.5% | — |
| Omni-MATH | 27.6% | — |
| LiveBench Math | 52.3% | — |
| MATH Level 5 | 56.9% | — |
| FrontierMath (Feb 2025 set) | 2.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
Claude 3.5 Sonnet: 28.6 (#245), DeepSeek-V2.5 (Sep 2024): 34.8 (#193)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Expert | 1265 | 1266 |
| GPQA Diamond | 55.3% | — |
| Humanity's Last Exam | 4.1% | — |
| MMLU-Pro | 77.7% | — |
| Confabulations | 19.9% | — |
| GPQA (HELM) | 56.5% | — |
| MMLU | 87.3% | — |
Multimodal Not comparable
Claude 3.5 Sonnet: 26.5 (#120), DeepSeek-V2.5 (Sep 2024): —
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Vision | 1125 | — |
| Video-MME | 60% | — |
| GeoBench | 62% | — |
| VPCT | 33% | — |
Multilingual Too close to call
Claude 3.5 Sonnet: 43.2 (#185), DeepSeek-V2.5 (Sep 2024): 42.5 (#193)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Non-English | 1283 | 1273 |
| LMArena Chinese | 1272 | 1318 |
| LMArena French | 1305 | 1289 |
| LMArena German | 1297 | 1258 |
| LMArena Japanese | 1234 | 1228 |
| LMArena Korean | 1200 | 1209 |
| LMArena Russian | 1306 | 1289 |
| LMArena Spanish | 1290 | 1248 |
Instruction Following Claude 3.5 Sonnet leads
Claude 3.5 Sonnet: 68.8 (#182), DeepSeek-V2.5 (Sep 2024): 67.5 (#194)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Instruction Following | 1297 | 1280 |
| LiveBench Instruction Following | 69.3% | — |
| IFEval | 85.5% | — |
Long Context Too close to call
Claude 3.5 Sonnet: 39.9 (#167), DeepSeek-V2.5 (Sep 2024): 39.5 (#174)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Longer Query | 1311 | 1301 |
Writing & Preference Claude 3.5 Sonnet leads
Claude 3.5 Sonnet: 52.9 (#164), DeepSeek-V2.5 (Sep 2024): 49.8 (#187)
| Benchmark | Claude 3.5 Sonnet | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Text | 1298 | 1294 |
| LMArena Creative Writing | 1292 | 1285 |
| LMArena Multi-Turn | 1326 | 1297 |
| Short-Story Creative Writing | 80.3% | — |
| EQ-Bench Creative Writing | 1451 | — |
| WildBench | 79.2% | — |
| LiveBench Language | 53.8% | — |
Frequently asked questions
Is Claude 3.5 Sonnet better than DeepSeek-V2.5 (Sep 2024)?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 34.6 on the Noometry Index.
Is Claude 3.5 Sonnet or DeepSeek-V2.5 (Sep 2024) better for coding?
Claude 3.5 Sonnet scores higher on coding benchmarks: 39.0 versus 31.7 in the Noometry coding category.
How many benchmarks do Claude 3.5 Sonnet and DeepSeek-V2.5 (Sep 2024) share?
22 benchmarks have published results for both models. Claude 3.5 Sonnet has 60 scored results on Noometry and DeepSeek-V2.5 (Sep 2024) has 22.