DeepSeek-V3.1-Terminus
| Released | 2025-09-22 |
|---|---|
| Price per 1M | Not listed |
| Overall score | 46.5 · rank 143 |
| Provider | DeepSeek |
| Weights | Open weights |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 46.5 | 143 | 4 |
| Overall: Coding | 44.8 | 158 | 3 |
| Overall: Reasoning | 47.5 | 136 | 2 |
| Overall: Agentic work | 45.5 | 150 | 3 |
| Overall: Frontend | 44.3 | 144 | 1 |
| Coding | 44.3 | 162 | 1 |
| Reasoning | 47.5 | 131 | 2 |
| General | 50.7 | 97 | 1 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 46.5 ±4.2 | 4 |
| Thinking | 50.7 ±7.0 | 1 |
Evidence
Coding 44.3 ±4.7
Reasoning 47.5 ±4.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| CritPt | 1.7% | Epoch AI | 2026-11-11 |
| Kagi Reasoning Benchmark | 57.4% | Kagi | 2026-11-11 |
General 50.7 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Text | 1,416 | LMArena | 2026-11-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| LMArena Text | 1,415 | LMArena | A stronger configuration result was used |
Names matched to this model
deepseek-v3-1-terminus