Google: palm-2
2 benchmarks
Leaderboard position
#1025
Overall rank (of 1121 models)
-209
Elo rating
2
Benchmarks
Top 91.4%
Percentile
In Google (of 72 models)
Above: Sarvam AI: sarvam-m #1024
Below: DeepSeek: deepseek-coder-v2-lite-instruct #1026
Benchmark highlights
Best benchmark
Text Arena (Snapshot from 20 june 2026)
1137
Nearest competitor from another company
The model from another company with the most similar benchmark scores.
appeared as nearest in 1 of 2 benchmarks
1/2
50.0%
1/2
50.0%
| Model | Votes | % of benchmarks |
|---|---|---|
| Snowflake: arctic-instruct | 1 | 50.0% |
| Meta Llama: llama-2-13b-chat | 1 | 50.0% |
Benchmarks
Predicted scores
Estimated from 25 similar models and the global Elo rating. These are projections, not measured results.
| Benchmark | Predicted | Confidence | Based on |
|---|---|---|---|
| MATH-500 | ~67.93 | 50% | — |
| AIME 2025 | ~31.21 | 50% | — |
| AIME 2026 | ~52.93 | 50% | — |
| HMMT Feb 2026 | ~44.36 | 50% | — |
| GPQA Diamond | ~34.69 | 50% | — |
| MMLU-Pro | ~35.13 | 50% | — |
| Humanity's Last Exam (no tools) | ~4 | 50% | — |
| IFEval | ~67.74 | 50% | — |
| IFBench | ~35.73 | 50% | — |
| HumanEval | ~42.83 | 50% | — |
| MBPP | ~36.59 | 50% | — |
| Terminal-Bench v2.1 | ~8.6 | 50% | — |
| MMLU | ~40.27 | 50% | — |
| SWE-bench Verified | ~5.83 | 50% | — |
| SWE-bench Pro | ~5.24 | 50% | — |
| SciCode | ~10.31 | 50% | — |
| AA-LCR | ~22.32 | 50% | — |
| Artificial Analysis Intelligence Index v5.0.1 | ~12 | 50% | — |
| Humanity's Last Exam (with tools) | ~11.9 | 50% | — |
| GSM8K | ~42.82 | 50% | — |
| -(CoderPPL) | ~-7.51 | 50% | — |
| LiveCodeBench v6 | ~31.64 | 50% | — |
| MMMU-Pro | ~32.99 | 50% | — |
| MATH | ~16.13 | 50% | — |
| BBH | ~41.73 | 50% | — |