Up to four models
Grok 4.20 vs Qwen3.8-27B
On the index (GPT-4 = 100), Grok 4.20 scores 156 and Qwen3.8-27B 155. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
Grok 4.20
156140–172
WeightsClosed
ReleasedFeb 2026
Price in / out$1.25 / $2.5
Context2M
Core tests taken4 of 9
155139–171
WeightsOpen
ReleasedAug 2026
Price in / out$0.060 / $2.2
Context1.02M
Core tests taken3 of 9
Grok 4.20 and Qwen3.8-27B are too close to call (1 point apart, within the uncertainty), ahead on 1 of 2 shared core tests, at 2.6× the price.
Test by testthird-party results
Grok 4.20
Qwen3.8-27B
Every result as a table
| Test | Grok 4.20 | Qwen3.8-27B |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Vals AI) | 88.6% | 88.9% |
| MMLU Pro | 86.3% | 84.3% |
| MMMU Pro | 83.5% | 83.9% |
| Coding | ||
| Code Migration | 0.3% | 14.2% |
| LiveCodeBench | 84.3% | 84.0% |
| SWE-bench | 72.2% | 86.0% |
| Vibe Code Bench v1.1 | 4.1% | 64.8% |
| Agents | ||
| DTBench | 90.1% | 88.0% |
| LMCA | 38.7% | 41.4% |
| Professional work | ||
| Finance Agent (v2) | 28.5% | 48.5% |
| EMB | 11.7% | 59.7% |
| LegalBench | 77.7% | 82.4% |
| Legal Research Bench | 13.9% | 36.1% |
| MedCode | 32.2% | 28.7% |
| MedScribe | 63.4% | 83.9% |
| MortgageTax | 45.4% | 64.9% |
| Tax Agent Bench | 7.0% | 30.9% |
| TaxEval v2 | 74.1% | 70.9% |
| Novel problems | ||
| ARC-AGI-2 | 65.1% | 42.4% |
| ARC-AGI-1 | 89.5% | 87.5% |
Tests at least two of these models have taken. Bold rows are the core tests.