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 | Qwen | Anthropic |
|---|---|---|---|---|
| Arena ELOHuman preference | 1,401 | Not in our dataset | 1,400 | 1,402 |
| LiveBench 1.0General reasoning | 69.6% | Not in our dataset | Not in our dataset | Not in our dataset |
| GPQA DiamondScientific knowledge | 82.8% | 73.0% | Not in our dataset | 76.1% |
| HLEAcademic reasoning | 24.3% | 9.7% | 15.8% | 7.8% |
| MMLU ProMulti-task understanding | Not in our dataset | Not in our dataset | 84.5% | Not in our dataset |
| SimpleBenchCommonsense reasoning | Not in our dataset | Not in our dataset | Not in our dataset | 45.5% |
| AIME 2025Mathematics | Not in our dataset | 80.7% | Not in our dataset | 70.5% |
| AiderCode editing | Not in our dataset | Not in our dataset | Not in our dataset | 61.3% |
| SWE-Bench VerifiedAgentic coding | Not in our dataset | 73.3% | Not in our dataset | 72.7% |
| LiveCodeBenchCompetitive coding | Not in our dataset | 53.2% | Not in our dataset | 55.9% |
| Tau-BenchAgentic tool use | Not in our dataset | 83.2% | 71.9% | 83.8% |
| ARC-AGI-2Visual reasoning | 5.7% | 4.0% | Not in our dataset | 5.9% |
Pricing & context Published USD rates per 1M tokens
| Specification | GPT-5 Nano Thinking (5.4) | Claude Haiku 4.5 Thinking | 3 235B A22B thinking-2507 | Claude Sonnet 4 4.0 Thinking |
|---|---|---|---|---|
| Input | $0.05 | $1 | $0.45 | $3 |
| Output | $0.4 | $5 | $3.5 | $15 |
| Context window | 400K | 200K | 128K | 200K |
| 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.