Model comparison
Claude 3 Haiku vs DeepSeek-V3.2-Exp
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 25.9 on the Noometry Index.
Last verified . 25 shared benchmarks.
DeepSeekSummary
- They share 25 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and DeepSeek-V3.2-Exp in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 17.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.8% for Claude 3 Haiku 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 Haiku | DeepSeek-V3.2-Exp | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 25.9 | 44.3 |
| Released | 2024-03-07 | 2025-09-29 |
| Weights | Proprietary | Open |
| Context window | — | 164K |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.26 |
| Output $ / M tokens | — | $0.38 |
| Results tracked | 37 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 26.4 (#325), DeepSeek-V3.2-Exp: 46.5 (#65)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| WeirdML | 9.8% | 39.5% |
| LMArena Coding | 1199 | 1454 |
| SWE-bench Verified (bash only) | — | 70% |
| Aider Polyglot | — | 74.2% |
| LMArena WebDev | — | 1362 |
| SWE-bench Multilingual | — | 59% |
| SciCode | — | 38.9% |
| BigCodeBench Instruct | 39.4% | — |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |
Agentic & Tool Use Not comparable
Claude 3 Haiku: —, DeepSeek-V3.2-Exp: 32.7 (#59)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| Terminal-Bench | — | 39.6% |
| APEX-Agents | — | 21.3% |
| Berkeley Function Calling Leaderboard | — | 56.7% |
| TheAgentCompany | — | 42.9% |
| Vending-Bench 2 | — | 1,034 |
Reasoning DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 16.3 (#307), DeepSeek-V3.2-Exp: 22.1 (#208)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| Kagi LLM Benchmark | 34.2% | 52.2% |
| LMArena Hard Prompts | 1174 | 1434 |
| DTBench | 50.1% | 87.7% |
| LMCA | 8.8% | 29.1% |
| Epoch Capabilities Index | 118.35 | 146.27 |
| ARC-AGI-2 | — | 4% |
| NYT Connections (extended) | — | 36.7% |
| ARC-AGI-1 | — | 57% |
| CritPt | — | 2.9% |
| Chess Puzzles | — | 14% |
| Thematic Generalization | — | 65% |
| ForecastBench | 53.2 | — |
| WinoGrande | 74.2% | — |
Math DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 9.8 (#319), DeepSeek-V3.2-Exp: 41.7 (#87)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 87.8% |
| LMArena Math | 1188 | 1435 |
| MathArena Final-Answer Competitions | — | 57.7% |
| ProofBench | — | 8% |
| MATH Level 5 | 14.9% | — |
| FrontierMath (Feb 2025 set) | — | 22.1% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 17.3 (#285), DeepSeek-V3.2-Exp: 51.7 (#66)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| GPQA Diamond | 36.3% | 83.4% |
| LMArena Expert | 1148 | 1436 |
| Confabulations | 34.2% | — |
| Vectara Hallucination Rate | — | 5.3% |
| MMLU | 73.8% | — |
Multimodal Not comparable
Claude 3 Haiku: 23.6 (#128), DeepSeek-V3.2-Exp: —
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |
Multilingual DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 36.0 (#243), DeepSeek-V3.2-Exp: 52.2 (#90)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Non-English | 1178 | 1409 |
| LMArena Chinese | 1155 | 1461 |
| LMArena French | 1195 | 1433 |
| LMArena German | 1174 | 1440 |
| LMArena Japanese | 1102 | 1374 |
| LMArena Korean | 1109 | 1371 |
| LMArena Russian | 1204 | 1424 |
| LMArena Spanish | 1166 | 1440 |
Instruction Following DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 61.3 (#247), DeepSeek-V3.2-Exp: 74.5 (#93)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Instruction Following | 1173 | 1413 |
Long Context DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 36.1 (#237), DeepSeek-V3.2-Exp: 47.6 (#16)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Longer Query | 1190 | 1428 |
| Fiction.LiveBench | — | 83.3% |
| CL-bench | — | 13.2% |
| CL-bench Life | — | 9.5% |
Writing & Preference DeepSeek-V3.2-Exp leads
Claude 3 Haiku: 29.7 (#291), DeepSeek-V3.2-Exp: 62.4 (#77)
| Benchmark | Claude 3 Haiku | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Text | 1195 | 1425 |
| LMArena Creative Writing | 1157 | 1403 |
| EQ-Bench Creative Writing | 717 | 1515 |
| LMArena Multi-Turn | 1190 | 1427 |
Frequently asked questions
Is Claude 3 Haiku better than DeepSeek-V3.2-Exp?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 25.9 on the Noometry Index.
Is Claude 3 Haiku or DeepSeek-V3.2-Exp better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 26.4 in the Noometry coding category.
How many benchmarks do Claude 3 Haiku and DeepSeek-V3.2-Exp share?
25 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and DeepSeek-V3.2-Exp has 49.