Model comparison
Claude 3.7 Sonnet vs DeepSeek-V3.2-Exp
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 39.5 on the Noometry Index.
Last verified . 28 shared benchmarks.
DeepSeekSummary
- They share 28 benchmarks with published results for both. Claude 3.7 Sonnet scores higher in 2 categories and DeepSeek-V3.2-Exp in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 39.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 57.8% for Claude 3.7 Sonnet and 87.8% for DeepSeek-V3.2-Exp.
- DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.
Side by side
| Claude 3.7 Sonnet | DeepSeek-V3.2-Exp | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 39.5 | 44.3 |
| Released | 2025-02-24 | 2025-09-29 |
| Weights | Proprietary | Open |
| Context window | — | 164K |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.26 |
| Output $ / M tokens | — | $0.38 |
| Results tracked | 58 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 40.6 (#136), DeepSeek-V3.2-Exp: 46.5 (#65)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| SWE-bench Verified (bash only) | 52.8% | 70% |
| Aider Polyglot | 64.9% | 74.2% |
| LMArena Coding | 1361 | 1454 |
| SWE-bench Verified | 61% | — |
| LMArena WebDev | — | 1362 |
| SWE-bench Multilingual | — | 59% |
| SciCode | — | 38.9% |
| GSO | 3.8% | — |
| WeirdML | — | 39.5% |
| LiveBench Coding | 74.5% | — |
| CadEval | 54% | — |
Agentic & Tool Use Claude 3.7 Sonnet leads
Claude 3.7 Sonnet: 34.1 (#50), DeepSeek-V3.2-Exp: 32.7 (#59)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| TheAgentCompany | 30.9% | 42.9% |
| Terminal-Bench | — | 39.6% |
| APEX-Agents | — | 21.3% |
| Berkeley Function Calling Leaderboard | — | 56.7% |
| Cybench | 20% | — |
| DeepResearch Bench | 43.6% | — |
| OSWorld | 35.8% | — |
| METR Time Horizons | 60% | — |
| Vending-Bench 2 | — | 1,034 |
Reasoning DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 18.6 (#277), DeepSeek-V3.2-Exp: 22.1 (#208)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| ARC-AGI-2 | 0.9% | 4% |
| ARC-AGI-1 | 28.6% | 57% |
| LMArena Hard Prompts | 1333 | 1434 |
| Epoch Capabilities Index | 141.16 | 146.27 |
| SimpleBench | 46.4% | — |
| Kagi LLM Benchmark | — | 52.2% |
| NYT Connections (extended) | — | 36.7% |
| CritPt | — | 2.9% |
| Chess Puzzles | — | 14% |
| EnigmaEval | 4.2% | — |
| Thematic Generalization | — | 65% |
| LiveBench Reasoning | 87.8% | — |
| DTBench | — | 87.7% |
| LiveBench Data Analysis | 74% | — |
| LMCA | — | 29.1% |
| ForecastBench | 61.8 | — |
| LiveBench | 76.1% | — |
Math DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 37.5 (#153), DeepSeek-V3.2-Exp: 41.7 (#87)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 57.8% | 87.8% |
| LMArena Math | 1337 | 1435 |
| FrontierMath (Feb 2025 set) | 4.1% | 22.1% |
| MathArena Final-Answer Competitions | — | 57.7% |
| ProofBench | — | 8% |
| Omni-MATH | 33% | — |
| LiveBench Math | 79% | — |
| MATH Level 5 | 91.2% | — |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 39.8 (#130), DeepSeek-V3.2-Exp: 51.7 (#66)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| GPQA Diamond | 79.7% | 83.4% |
| LMArena Expert | 1321 | 1436 |
| Humanity's Last Exam | 8% | — |
| MMLU-Pro | 78.4% | — |
| Confabulations | 14.7% | — |
| Vectara Hallucination Rate | — | 5.3% |
| GPQA (HELM) | 60.8% | — |
Multimodal Not comparable
Claude 3.7 Sonnet: 33.7 (#95), DeepSeek-V3.2-Exp: —
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Vision | 1169 | — |
| GeoBench | 68% | — |
| VPCT | 39% | — |
| SpatialViz-Bench | 33.9% | — |
Multilingual DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 44.1 (#179), DeepSeek-V3.2-Exp: 52.2 (#90)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Non-English | 1296 | 1409 |
| LMArena Chinese | 1299 | 1461 |
| LMArena French | 1303 | 1433 |
| LMArena German | 1301 | 1440 |
| LMArena Japanese | 1267 | 1374 |
| LMArena Korean | 1249 | 1371 |
| LMArena Russian | 1311 | 1424 |
| LMArena Spanish | 1298 | 1440 |
Instruction Following DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 72.9 (#125), DeepSeek-V3.2-Exp: 74.5 (#93)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Instruction Following | 1352 | 1413 |
| LiveBench Instruction Following | 81.3% | — |
| IFEval | 83.4% | — |
Long Context Claude 3.7 Sonnet leads
Claude 3.7 Sonnet: 50.3 (#10), DeepSeek-V3.2-Exp: 47.6 (#16)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| Fiction.LiveBench | 83.3% | 83.3% |
| LMArena Longer Query | 1373 | 1428 |
| CL-bench | — | 13.2% |
| CL-bench Life | — | 9.5% |
Writing & Preference DeepSeek-V3.2-Exp leads
Claude 3.7 Sonnet: 54.4 (#150), DeepSeek-V3.2-Exp: 62.4 (#77)
| Benchmark | Claude 3.7 Sonnet | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Text | 1314 | 1425 |
| LMArena Creative Writing | 1332 | 1403 |
| EQ-Bench Creative Writing | 1412 | 1515 |
| LMArena Multi-Turn | 1339 | 1427 |
| Short-Story Creative Writing | 81.1% | — |
| WildBench | 81.4% | — |
| LiveBench Language | 59.9% | — |
Frequently asked questions
Is Claude 3.7 Sonnet better than DeepSeek-V3.2-Exp?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 39.5 on the Noometry Index.
Is Claude 3.7 Sonnet or DeepSeek-V3.2-Exp better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 40.6 in the Noometry coding category.
How many benchmarks do Claude 3.7 Sonnet and DeepSeek-V3.2-Exp share?
28 benchmarks have published results for both models. Claude 3.7 Sonnet has 58 scored results on Noometry and DeepSeek-V3.2-Exp has 49.