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 | OpenAI | Anthropic | OpenAI | Zhipu AI |
|---|---|---|---|---|
| Arena ELOHuman preference | Not in our dataset | 1,462 | 1,467 | 1,465 |
| LiveBench 1.0General reasoning | 74.6% | 76.0% | 77.9% | Not in our dataset |
| GPQA DiamondScientific knowledge | 92.4% | Not in our dataset | Not in our dataset | 86.2% |
| HLEAcademic reasoning | 35.4% | Not in our dataset | Not in our dataset | 31.0% |
| MMLU ProMulti-task understanding | 87.4% | Not in our dataset | Not in our dataset | 85.9% |
| SimpleBenchCommonsense reasoning | 45.8% | 60.6% | 48.9% | 55.1% |
| AIME 2025Mathematics | 100.0% | Not in our dataset | Not in our dataset | Not in our dataset |
| GSOGraduate science | 27.4% | 37.3% | Not in our dataset | Not in our dataset |
| SWE-Bench VerifiedAgentic coding | 80.0% | Not in our dataset | Not in our dataset | 73.8% |
| Tau-BenchAgentic tool use | 82.0% | Not in our dataset | Not in our dataset | 84.1% |
| Terminal-BenchAgentic terminal | 54.0% | Not in our dataset | Not in our dataset | 63.5% |
| ARC-AGI-2Visual reasoning | 52.9% | Not in our dataset | 74.2% | Not in our dataset |
Pricing & context Published USD rates per 1M tokens
| Specification | GPT-5.2 Thinking | Claude Sonnet 5 Thinking | GPT-5.6 Terra Thinking | GLM-5.1 Thinking |
|---|---|---|---|---|
| Input | $1.75 | $2 | $2 | $1.4 |
| Output | $14 | $10 | $12 | $4.4 |
| Context window | 400K | 1.0M | 1.1M | 205K |
| 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.