Compare models. Find the differences.
Add, remove, or change models in the table.
Benchmark snapshot
3 benchmarks in common.
4 models selected. 3 benchmarks in common.
Reported scores; test settings may differ.
| Benchmark | xAI | Google | Anthropic | xAI |
|---|---|---|---|---|
| Arena ELOHuman preference | 1,430 | 1,438 | 1,456 | 1,411 |
| GPQA DiamondScientific knowledge | 84.3% | 82.3% | 83.4% | 87.5% |
| HLEAcademic reasoning | 17.6% | 8.7% | 17.7% | 25.4% |
| MMLU ProMulti-task understanding | Not in our dataset | 82.6% | 87.5% | 86.6% |
| SimpleBenchCommonsense reasoning | Not in our dataset | Not in our dataset | 54.3% | 60.5% |
| AIME 2025Mathematics | Not in our dataset | Not in our dataset | 87.0% | 91.7% |
| GSOGraduate science | Not in our dataset | Not in our dataset | 14.7% | Not in our dataset |
| AiderCode editing | Not in our dataset | Not in our dataset | Not in our dataset | 79.6% |
| SWE-Bench VerifiedAgentic coding | Not in our dataset | Not in our dataset | 77.2% | Not in our dataset |
| LiveCodeBenchCompetitive coding | 76.5% | 77.1% | Not in our dataset | Not in our dataset |
| Tau-BenchAgentic tool use | Not in our dataset | Not in our dataset | 86.2% | 76.5% |
| Terminal-BenchAgentic terminal | Not in our dataset | Not in our dataset | 42.8% | 23.1% |
| ARC-AGI-2Visual reasoning | 5.3% | Not in our dataset | 13.6% | 16.0% |
Pricing & context Published USD rates per 1M tokens
| Specification | Grok 4 Fast Thinking | Gemma 4 26B A4B IT Thinking | Claude Sonnet 4 4.5 Thinking | Grok 4 |
|---|---|---|---|---|
| Input | $0.2 | $0.095 | $3 | $3 |
| Output | $0.5 | $0.284 | $15 | $15 |
| Context window | 2.0M | 262K | 1.0M | 256K |
| Pricing details | Explore pricing → | Explore pricing → | Explore pricing → | Explore pricing → |
How to read this comparison
Each column keeps the exact model version you selected. Shared benchmarks are the intersection of their reported results; “all coverage” also shows where evidence is missing. Missing results never count as zero.
These are published benchmark scores, not a controlled head-to-head test. The available evidence does not establish matching evaluation settings, so we show no cross-model winner highlights or deltas. Composite quality scores are omitted because they can summarize different benchmark sets.
Scores and statistics are computed upstream. This page reads a fixed data snapshot; its date describes the snapshot, not the date of every measurement.