Skip to content

The Aggregate

Unified LLM rankings, daily

Models What's New Trends Benchmarks Skill Maps
How It Works Guesswork Metrics
The Aggregate
Loading data...

SAFIM — leaderboard

Metric: Average. Source: huggingface.co. 15 models tracked.

Top models

#ModelScore
1deepseek-coder-33B-base69.01
2deepseek-coder-6.7B-base63.4
3GPT-4 Preview (1106)53.28
4CodeLlama-13B-hf52.78
5deepseek-coder-1.3B-base52.63
6CodeLlama-34B-hf49.66
7CodeLlama-7B-hf45
8GPT-3.5 Turbo (0301)40.86

Interactive version: theaggregate.ai/benchmark?slug=safim · How the rankings work · Data refreshed daily, snapshot 2026-07-22.

Built by Mikhail Doroshenko — AI researcher, co-author of Humanity’s Last Exam. Independent project; no affiliation with any AI lab.