Compare models. Find the differences.
Add, remove, or change models in the table.
Benchmark snapshot
1 benchmarks in common.
4 models selected. 1 benchmarks in common.
Reported scores; test settings may differ.
| Benchmark | Anthropic | xAI | Qwen | Moonshot |
|---|---|---|---|---|
| Arena ELOHuman preference | 1,402 | 1,442 | 1,400 | 1,418 |
| LiveBench 1.0General reasoning | Not in our dataset | 62.3% | Not in our dataset | Not in our dataset |
| GPQA DiamondScientific knowledge | 76.1% | Not in our dataset | Not in our dataset | 74.2% |
| HLEAcademic reasoning | 7.8% | Not in our dataset | 15.8% | Not in our dataset |
| MMLU ProMulti-task understanding | Not in our dataset | Not in our dataset | 84.5% | 81.9% |
| SimpleBenchCommonsense reasoning | 45.5% | Not in our dataset | Not in our dataset | Not in our dataset |
| AIME 2025Mathematics | 70.5% | Not in our dataset | Not in our dataset | 51.0% |
| AiderCode editing | 61.3% | Not in our dataset | Not in our dataset | Not in our dataset |
| SWE-Bench VerifiedAgentic coding | 72.7% | Not in our dataset | Not in our dataset | 69.2% |
| LiveCodeBenchCompetitive coding | 55.9% | Not in our dataset | Not in our dataset | Not in our dataset |
| Tau-BenchAgentic tool use | 83.8% | Not in our dataset | 71.9% | Not in our dataset |
| ARC-AGI-2Visual reasoning | 5.9% | Not in our dataset | Not in our dataset | Not in our dataset |
Pricing & context Published USD rates per 1M tokens
| Specification | Claude Sonnet 4 4.0 Thinking | Grok 4.3 | 3 235B A22B thinking-2507 | Kimi K2 0905 Preview |
|---|---|---|---|---|
| Input | $3 | $1.25 | $0.45 | $0.6 |
| Output | $15 | $2.5 | $3.5 | $2.5 |
| Context window | 200K | 1.0M | 128K | 262K |
| 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.