Models · Compare
Claude 3.5 Sonnet vs DeepSeek-Coder-V2 Instruct
Rows highlighted in warm gray are where the models differ. Numbers carry their as-of date and primary source.
Specs
| Field | A: Claude 3.5 Sonnet | B: DeepSeek-Coder-V2 Instruct |
|---|---|---|
| Released | 2024-06-20 | 2024-06-17 |
| Developer | Anthropic | DeepSeek |
| Openness | Proprietary | Open weights |
| License | Proprietary | DeepSeek License |
| OSI-approved | no | no |
| Data released | no | no |
| Training code | no | no |
| Architecture | unknown | moe |
| Total params | — | 236B |
| Active params | — | 21B |
| Experts | — | — |
| Context window | 200K | 128K |
| Attention | unknown | mla |
| Position enc. | unknown | rope-yarn |
| Pretraining tokens | — | 6.0T |
| Post-training | rlhf, constitutional | sft, rlhf |
| Training hardware | — | — |
| $/M input | — | — |
| $/M output | — | — |
| Output tok/sec | — | — |
Benchmarks
Missing scores render as not reported; never inferred. Bold highlights the leader per benchmark.
General reasoning
Code
| HumanEval | — | 90.2 2024-06-17 |
| LiveCodeBench | — | 43.4 2024-06-17 |
Context · A
The first Claude release to beat its own larger sibling (Claude 3 Opus) on most benchmarks. Established Artifacts, driving a wave of code-and-canvas product copies.
Context · B
Code specialist branch of the V2 line. Continued pretraining from an intermediate V2 checkpoint on an additional 6 trillion code and math tokens, expanding programming language coverage from 86 to 338 and extending context from 16K to 128K. The Instruct variant reported HumanEval scores competitive with GPT-4 Turbo.
Claude 3.5 Sonnet detail → · DeepSeek-Coder-V2 Instruct detail →