0 vs 1 benchmarks won
Anthropic Claude 3 Haiku | DeepSeek DeepSeek-V4-Pro | |
|---|---|---|
| Overview | ||
| Company | Anthropic | DeepSeek |
| Release date | Mar 4 2024 | Apr 24 2026 |
| Model type | — | — |
| Open source | No | Yes |
| Specifications | ||
Parameters | — | — |
Context window | — | — |
| Benchmarks | ||
Science reasoning GPQA Diamond | 33.3% | 90.1% |
Software engineering SWE-Bench Verified | — | — |
Multimodal understanding MMMU | — | — |
| Timeline | ||
| Release gap | Claude 3 Haiku shipped 781 days before DeepSeek-V4-Pro | |
DeepSeek-V4-Pro leads Claude 3 Haiku on 1 of the tracked benchmarks (GPQA Diamond, SWE-Bench Verified, MMMU). Claude 3 Haiku shipped 781 days before DeepSeek-V4-Pro, so benchmark comparisons should account for the intervening progress.
DeepSeek-V4-Pro is an open-source / open-weight model; Claude 3 Haiku is proprietary.
On GPQA Diamond, DeepSeek-V4-Pro scores 90.1%, 56.8 points above Claude 3 Haiku at 33.3%.