Doubts Over Soofi's "Best Open-Source Model" Claim
JJitsev · x · 2026-07-16
The author questions Soofi's claim as the "strongest open-source model." The rationale is that while its training process was similar to the public version of Nemotron-3-Nano, the dataset used included all rewritten benchmark sets.
This implies the comparison is likely tainted by benchmark leakage/contamination, making the conclusions highly unconvincing.
Related event: SOOFI benchmark claims challenged over leakage and baseline reporting(13 posts)→
More from Models
- Unreleased 'GPT 6 Sol' model spotted in OpenAI's API — SteveEricJordan · 2026-09-11
- DeepSeek update keeps cache hits mid-conversation, cuts costs 36.6% — teortaxesTex · 2026-09-11
- AI Sextet offers 6 models free and unlimited for 14 days, including DeepSeek and Qwen — airesearch12 · 2026-09-11
- BullshitBench update: GPT-6-Astra beats all prior OpenAI models but still trails Anthropic — scaling01 · 2026-09-11
- Astra Scores 83% on GauntletBench, First Computer-Use Agent to Beat Human Baseline — ducha_aiki · 2026-09-11
- Kimi K2.8 Preview rolls out: near-K3 coding performance, 1M context for all tiers — teortaxesTex · 2026-09-11