Model comparison
Claude 3 Haiku vs DeepSeek-R1
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 25.9 on the Noometry Index.
Last verified . 26 shared benchmarks.
DeepSeekSummary
- They share 26 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and DeepSeek-R1 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 43.8 to 9.8.
- The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 96.6% for DeepSeek-R1.
Side by side
| Claude 3 Haiku | DeepSeek-R1 | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 25.9 | 42.3 |
| Released | 2024-03-07 | 2025-01-20 |
| Weights | Proprietary | Proprietary |
| Context window | — | 164K |
| Max output | — | 64K |
| Input $ / M tokens | — | $0.50 |
| Output $ / M tokens | — | $2.15 |
| Results tracked | 37 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
Claude 3 Haiku: 26.4 (#325), DeepSeek-R1: 46.3 (#68)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| WeirdML | 9.8% | 41.6% |
| LMArena Coding | 1199 | 1427 |
| Aider Polyglot | — | 71.4% |
| SciCode | — | 35.7% |
| BigCodeBench Instruct | 39.4% | — |
| LiveBench Coding | — | 66.7% |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| ALE-Bench | — | 804.12 |
| AlgoTune | — | 1.7 |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |
Agentic & Tool Use Not comparable
Claude 3 Haiku: —, DeepSeek-R1: 30.7 (#75)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| DeepResearch Bench | — | 35.1% |
| BALROG | — | 34.9% |
| METR Time Horizons | — | 53.8% |
Reasoning DeepSeek-R1 leads
Claude 3 Haiku: 16.3 (#307), DeepSeek-R1: 18.6 (#278)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| Kagi LLM Benchmark | 34.2% | 69.4% |
| LMArena Hard Prompts | 1174 | 1416 |
| Epoch Capabilities Index | 118.35 | 141.29 |
| ForecastBench | 53.2 | 60 |
| ARC-AGI-2 | — | 1.3% |
| SimpleBench | — | 40.8% |
| ARC-AGI-1 | — | 21.2% |
| CritPt | — | 1.1% |
| LiveBench Reasoning | — | 83.2% |
| DTBench | 50.1% | — |
| LiveBench Data Analysis | — | 69.8% |
| LMCA | 8.8% | — |
| LiveBench | — | 71.6% |
| WinoGrande | 74.2% | — |
Math DeepSeek-R1 leads
Claude 3 Haiku: 9.8 (#319), DeepSeek-R1: 43.8 (#79)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 66.4% |
| LMArena Math | 1188 | 1400 |
| MATH Level 5 | 14.9% | 96.6% |
| Omni-MATH | — | 42.4% |
| LiveBench Math | — | 80.7% |
Knowledge DeepSeek-R1 leads
Claude 3 Haiku: 17.3 (#285), DeepSeek-R1: 44.5 (#87)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| GPQA Diamond | 36.3% | 76.3% |
| Confabulations | 34.2% | 12.7% |
| LMArena Expert | 1148 | 1394 |
| MMLU-Pro | — | 79.3% |
| Vectara Hallucination Rate | — | 11.3% |
| GPQA (HELM) | — | 66.6% |
| MMLU | 73.8% | — |
Multimodal Not comparable
Claude 3 Haiku: 23.6 (#128), DeepSeek-R1: —
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |
Multilingual DeepSeek-R1 leads
Claude 3 Haiku: 36.0 (#243), DeepSeek-R1: 52.4 (#85)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| LMArena Non-English | 1178 | 1412 |
| LMArena Chinese | 1155 | 1442 |
| LMArena French | 1195 | 1417 |
| LMArena German | 1174 | 1404 |
| LMArena Japanese | 1102 | 1391 |
| LMArena Korean | 1109 | 1360 |
| LMArena Russian | 1204 | 1423 |
| LMArena Spanish | 1166 | 1411 |
Instruction Following DeepSeek-R1 leads
Claude 3 Haiku: 61.3 (#247), DeepSeek-R1: 72.0 (#143)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| LMArena Instruction Following | 1173 | 1382 |
| LiveBench Instruction Following | — | 80.5% |
| IFEval | — | 78.4% |
Long Context DeepSeek-R1 leads
Claude 3 Haiku: 36.1 (#237), DeepSeek-R1: 45.4 (#36)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| LMArena Longer Query | 1190 | 1391 |
| Fiction.LiveBench | — | 75% |
Writing & Preference DeepSeek-R1 leads
Claude 3 Haiku: 29.7 (#291), DeepSeek-R1: 61.4 (#88)
| Benchmark | Claude 3 Haiku | DeepSeek-R1 |
|---|---|---|
| LMArena Text | 1195 | 1428 |
| LMArena Creative Writing | 1157 | 1405 |
| EQ-Bench Creative Writing | 717 | 1500 |
| LMArena Multi-Turn | 1190 | 1405 |
| Short-Story Creative Writing | — | 83% |
| WildBench | — | 82.8% |
| LiveBench Language | — | 48.5% |
Frequently asked questions
Is Claude 3 Haiku better than DeepSeek-R1?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 25.9 on the Noometry Index.
Is Claude 3 Haiku or DeepSeek-R1 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 26.4 in the Noometry coding category.
How many benchmarks do Claude 3 Haiku and DeepSeek-R1 share?
26 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and DeepSeek-R1 has 52.