SphinX Benchmark Results — The Aggregate

SphinX — benchmark results

Alpha-VLLM's SPHINX multimodal model on a LLaMA-2-13B backbone, jointly mixing weights, tasks and visual embeddings for vision-language work (November 2023). Provider: Other. Released 2023-11-02. Access: Open.

Unified ELO 1472 ± 42, rank #903 of 1776 rated models, from 7 benchmark results.

Strongest benchmark results

Benchmark	Score	Metric	Percentile
Open LLM Leaderboard - MATH Level 5	30.82	Score	82.3
Open LLM Leaderboard - MMLU-Pro	37.4	Score	79.7
Open LLM Leaderboard - MuSR	12.7	Score	69.6
Open LLM Leaderboard - BBH	34.71	Score	69.3
Open LLM Leaderboard - IFEval	57.25	Score	69.2
Open LLM Leaderboard - GPQA	6.38	Score	53.5
MMMU Benchmark	32.9	Validation Score	7.7

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