Muse Spark 1.1 IQ Ranking Rises by Reducing Hallucinations
ArtificialAnlys · x · 2026-07-11
Artificial Analysis observed that Muse Spark 1.1's ranking improvement pattern on the Intelligence Index is the opposite of Grok 4.5. Grok 4.5's improvement came with a simultaneous rise in both accuracy and hallucination rates, whereas Muse Spark 1.1's jump from 4th to 18th place is primarily due to an "abstention" mechanism: with accuracy remaining basically flat, its hallucination rate dropped significantly by 35 percentage points to 38%.
Additionally, when running the Intelligence Index tests, Muse Spark 1.1 demonstrated better token efficiency, consuming 94 million output tokens, which is lower than similarly-scoring models like GPT-5.4 (xhigh) and GLM-5.2 (max).
More from Models
- Kimi K3 Is 'Open' Until You Succeed: License Caps Free Use at $20M Revenue — alex_verem · 2026-10-27
- Screenshot shows 'GPT-6 Astra Ultra' in ChatGPT model picker — unverified leak or joke — justalexoki · 2026-10-27
- DeepSeek Safety Eval Finds "Over-Attribution": Model Mistakes Ordinary Tasks for Tests — Much_Boysenberry9124 · 2026-10-27
- Cohere flags popular open-source model attempting to upload probe file to Beijing bucket — stanislavfort · 2026-10-27
- MissionReport benchmark: frontier models score just ~60% on government status reporting — Exp_Mark · 2026-10-27
- Anthropic's follow-up on Claude's unauthorized web actions: self-policing fails — AccBalanced · 2026-10-27