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 | Google | Anthropic | Google | Meta |
|---|---|---|---|---|
| Arena ELOHuman preference | 1,342 | 1,325 | 1,365 | 1,327 |
| LiveBench 1.0General reasoning | 43.7% | Not in our dataset | 49.2% | 59.5% |
| GPQA DiamondScientific knowledge | 40.9% | Not in our dataset | 42.4% | 69.8% |
| HLEAcademic reasoning | Not in our dataset | Not in our dataset | Not in our dataset | 5.7% |
| MMLU ProMulti-task understanding | 60.6% | Not in our dataset | 67.5% | 80.5% |
| SimpleBenchCommonsense reasoning | Not in our dataset | Not in our dataset | Not in our dataset | 27.7% |
| AIME 2025Mathematics | 18.8% | Not in our dataset | 24.0% | 15.9% |
| AiderCode editing | Not in our dataset | 28.0% | 4.9% | 15.6% |
| LiveCodeBenchCompetitive coding | 24.6% | Not in our dataset | 29.7% | 37.2% |
| ARC-AGI-2Visual reasoning | Not in our dataset | Not in our dataset | Not in our dataset | 0.0% |
Pricing & context Published USD rates per 1M tokens
| Specification | Gemma 3 12B IT Non-thinking | Claude 3.5 Haiku Non-thinking | Gemma 3 27B IT Non-thinking | Llama 4 Maverick Default (Non-thinking) |
|---|---|---|---|---|
| Input | $0.05 | $1 | $0.08 | $0.188 |
| Output | $0.15 | $5 | $0.35 | $0.652 |
| Context window | 131K | 200K | 131K | 524K |
| 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.