Comparison of Top Media Models: Open-Source Takes on Proprietary Systems
benchmarks open-source
| Source: Mastodon | Original article
Open-source image editing models trail proprietary ones. FLUX.2 ranks 14th, 78 ELO points behind Riverflow 2.0.
The media model leaderboard has sparked interest in the comparison between open-source and proprietary models, particularly in image editing. As of the latest update, the best open-source model, FLUX.2, trails behind Riverflow 2.0, a proprietary model, by 78 ELO points. This gap is notable, but the landscape is shifting in favor of open-source models.
The performance difference between open-source and proprietary models is narrowing. According to a 2025 benchmark analysis, the best open-source model, MiniMax-M2, is only 7 points behind the best proprietary model, GPT-5. This gap has decreased significantly from 15-20 points in 2024. Furthermore, open-source models offer a substantial cost advantage, with an average cost of $0.83 per million tokens compared to $6.03 for proprietary models.
As the open-source community continues to close the performance gap, it will be interesting to watch how proprietary models respond. The open LLM leaderboard will likely be a key indicator of the progress made by open-source models. With the leaderboard updated hourly, developers and users can track the latest rankings and capabilities of open-source models, making it an essential resource for those interested in the evolving AI landscape.
Sources
Back to AIPULSEN