DeepSeek-V3.2-Exp Thinking
| Released | 2025-09-29 |
|---|---|
| Price per 1M | Not listed |
| Overall score | 50.0 · rank 267 |
| Provider | DeepSeek |
| Weights | Open weights |
| Configuration | Thinking |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 50.0 | 267 | 4 |
| Overall: Coding | 49.1 | 273 | 3 |
| Overall: Reasoning | 50.8 | 256 | 1 |
| Overall: Agentic work | 49.5 | 277 | 3 |
| Overall: Frontend | 48.7 | 233 | 2 |
| Coding | 48.7 | 258 | 2 |
| Reasoning | 50.8 | 251 | 1 |
| General | 52.0 | 140 | 1 |
Evidence
Coding 48.7 ±1.5
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Aider Polyglot | 74.2% | Epoch AI | 2026-11-11 |
| SciCode | 38.9% | Epoch AI | 2026-11-11 |
Reasoning 50.8 ±5.3
General 52.0 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Text | 1,425 | LMArena | 2026-11-09 |
Names matched to this model
deepseek-v3-2-exp-thinking