Up to four models
DeepSeek-V4-Pro vs Grok 4.7
On the index (GPT-4 = 100), DeepSeek-V4-Pro scores 169 and Grok 4.7 168. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
DeepSeek-V4-Pro169152–185
WeightsOpen
ReleasedAug 2026
Price in / out$1.32 / $3.96
Context1.05M
Core tests taken6 of 9
Grok 4.7
168151–184
WeightsClosed
ReleasedSep 2026
Price in / out$2.0 / $6.0
Context500K
Core tests taken6 of 9
DeepSeek-V4-Pro and Grok 4.7 are too close to call (1 point apart, within the uncertainty), ahead on 1 of 6 shared core tests, and costs 1.5× less.
Test by testthird-party results
DeepSeek-V4-ProGrok 4.7Every result as a table
| Test | DeepSeek-V4-Pro | Grok 4.7 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 91.7% | 92.7% |
| SimpleQA Verified | 52.9% | 56.0% |
| Maths | ||
| FrontierMath (tiers 1-3) | 64.6% | 53.0% |
| AIME-style maths (OTIS mock) | 98.6% | 98.1% |
| ProofBench v1.1 | 50.0% | 26.0% |
| Chess puzzles | 46.0% | 41.0% |
| FrontierMath (tier 4) | 26.8% | 17.1% |
| Coding | ||
| Terminal-Bench 4.0 | 14.1% | 28.8% |
| Code Migration | 41.5% | 44.8% |
| CyberBench v1.1 | 72.9% | 69.5% |
| IOI | 51.6% | 57.7% |
| ProgramBench | 0.0% | 0.5% |
| Vibe Code Bench v1.1 | 82.3% | 86.2% |
| Agents | ||
| APEX-Agents | 47.3% | 54.6% |
| SRE Bench | 1.9% | 3.8% |
| Terminal-Bench-Science | 4.3% | 10.0% |
| DTBench | 93.9% | 96.0% |
| LMCA | 45.5% | 49.4% |
| Mystery game puzzles | 43.0% | 29.0% |
| Professional work | ||
| Finance Agent (v2) | 50.4% | 52.3% |
| EMB | 52.8% | 67.0% |
| HLAB-Vals-1.1 | 10.8% | 18.3% |
| LegalBench | 82.4% | 84.4% |
| Legal Research Bench | 40.9% | 47.1% |
| MedCode | 42.5% | 49.5% |
| MedScribe | 80.2% | 89.4% |
| Tax Agent Bench | 26.8% | 34.2% |
| Novel problems | ||
| ARC-AGI-2 | 61.3% | 61.4% |
| ARC-AGI-1 | 90.5% | 90.2% |
Tests at least two of these models have taken. Bold rows are the core tests.
