DeepSeek V4 Pro 0813
| Released | 2026-08-13 |
|---|---|
| Price per 1M | $0.66 in · $1.98 out |
| Overall score | 59.0 · rank 37 |
| Context window | 1,000,000 tokens |
| Provider | DeepSeek |
| Weights | Open weights (MIT) |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 59.0 | 37 | 34 |
| Overall: Coding | 58.9 | 40 | 26 |
| Overall: Reasoning | 59.6 | 39 | 13 |
| Overall: Agentic work | 58.9 | 45 | 25 |
| Overall: Frontend | 59.2 | 29 | 9 |
| Coding | 58.3 | 43 | 8 |
| Agents | 58.5 | 47 | 11 |
| Reasoning | 61.3 | 35 | 6 |
| Math | 58.7 | 41 | 4 |
| Knowledge | 57.2 | 47 | 3 |
| Frontend | 59.7 | 24 | 1 |
| General | 59.8 | 32 | 1 |
| Value vs price | +11.2 | 24 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 58.1 ±4.7 | 2 |
| High | 57.9 ±3.4 | 4 |
| Low | 60.3 ±5.7 | 2 |
| Max | 59.2 ±1.3 | 31 |
| Non-reasoning | 44.9 ±4.9 | 4 |
Evidence
Coding 58.3 ±2.5
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierCode | 28.6% | Epoch AI | 2026-11-11 |
| Code Migration v1.1 | 52.0% | Vals AI | 2026-11-09 |
| MirrorCode | 33.3% | Epoch AI | 2026-10-20 |
| IOI | 51.6% | Vals AI | 2026-10-29 |
| SWE-bench Verified (Vals run) | 96.4% | Vals AI | 2026-09-01 |
| SciCode | 51.0% | Epoch AI | 2026-11-11 |
| LiveCodeBench (Vals run) | 87.5% | Vals AI | 2026-09-01 |
| ProgramBench | 0.0% | Vals AI | 2026-11-09 |
Agents 58.5 ±1.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 4.0 (Vals run) | 14.1% | Vals AI | 2026-11-09 |
| SRE Bench | 1.9% | Vals AI | 2026-11-05 |
| Terminal-Bench-Science | 4.3% | Vals AI | 2026-11-09 |
| APEX-Agents | 47.3% | Epoch AI | 2026-11-11 |
| Terminal-Bench 2.1 (Vals run) | 54.7% | Vals AI | 2026-10-01 |
| EMB v1.1 (Excel financial modeling) | 65.1% | Vals AI | 2026-11-09 |
| Tax Agent Bench | 26.8% | Vals AI | 2026-11-09 |
| Finance Agent v2.1 | 57.1% | Vals AI | 2026-11-09 |
| SkillsBench | 53.8% | Vals AI | 2026-10-29 |
| HLAB-Vals 1.1 (legal agent) | 10.8% | Vals AI | 2026-11-09 |
| CyberBench v1.1 | 72.9% | Vals AI | 2026-11-05 |
Reasoning 61.3 ±2.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| ARC-AGI-2 | 61.3% | Epoch AI | 2026-11-11 |
| ARC-AGI-2 (ARC Prize verified) | 61.3% | ARC Prize | 2026-11-11 |
| CritPt | 18.0% | Epoch AI | 2026-11-11 |
| ARC-AGI-1 (ARC Prize verified) | 90.5% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 | 90.5% | Epoch AI | 2026-11-11 |
| GPQA Diamond | 91.7% | Epoch AI | 2026-08-18 |
| GPQA Diamond (Vals run) | 92.4% | Vals AI | 2026-09-01 |
| Chess Puzzles | 46.0% | Epoch AI | 2026-10-13 |
| Mystery Game Puzzles | 43.0% | Epoch AI | 2026-08-19 |
Math 58.7 ±1.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tiers 1-3 | 64.6% | Epoch AI | 2026-08-19 |
| FrontierMath Tier 4 | 26.8% | Epoch AI | 2026-08-19 |
| OTIS Mock AIME 2024-2025 | 98.6% | Epoch AI | 2026-08-18 |
| ProofBench v1.1 | 50.0% | Vals AI | 2026-11-09 |
| ProofBench | 50.0% | Epoch AI | 2026-11-11 |
Knowledge 57.2 ±3.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SimpleQA Verified | 52.9% | Epoch AI | 2026-08-27 |
| MMLU-Pro (Vals run) | 87.0% | Vals AI | 2026-09-01 |
| Legal Research Bench | 40.9% | Vals AI | 2026-11-09 |
Frontend 59.7 ±0.8
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Vibe Code Bench v1.1 | 82.3% | Vals AI | 2026-11-09 |
General 59.8 ±3.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LLM Debate Benchmark | 1,565 | Lech Mazur | 2026-10-26 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| ARC-AGI-1 | 90.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 13.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 87.2% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 90.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 13.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 87.2% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 | 0.8% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 59.7% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 56.3% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 0.8% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 59.7% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 56.3% | ARC Prize | A stronger configuration result was used |
| CritPt | 0.0% | Epoch AI | A stronger configuration result was used |
| SciCode | 39.9% | Epoch AI | A stronger configuration result was used |
Names matched to this model
deepseek-v4-pro-0813