‹ PublicAI Index
The LLM benchmark aggregator.
Muse Spark 1.1
Meta
Strongest in Computer use (#3 of 68, on 1 of its 2 boards), weakest in Toxicity avoidance (#156 of 268). Above par in 22 of 25 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of Gemini 3.1 Pro and GPT-6 Sol.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Safety59.8−1.6#7/3291/3
Jailbreak resistance63.6−1.4#6/2681/1
Harm refusal59.9−1.6#9/2921/2
Fairness67.9−2.1#16/2921/2
Secure code56.9−9.3#97/2701/1
Toxicity avoidance52.6−6.4#156/2681/1
Professional58−4.4#10/1941/2
Finance61.3−1.1#4/1621/1
Medical57.5−9#26/1491/1
Legal55.3−10.6#45/1731/1
Human preference65.4−3.4#12/3541/1
Human preference65.4−3.4#12/3541/1
Coding54.7−11.5#39/1792/6
Agentic coding54.9−12#40/1732/5
Agents56.1−10.3#46/2983/6
Computer use65.7−2.4#3/681/2
Knowledge work52.8−17.9#79/2082/2
Knowledge53.9−8.6#46/1971/3
Academic knowledge59.4−4.5#17/1231/1
Reasoning52.2−13.2#76/1952/5
Science60.5−3#23/1221/1
Mathematics47.2−16.6#97/1521/2
Core abilities49.2−16.1#108/1991/4
Instruction following52.9−20.9#30/621/1
Language45−21.6#46/621/1
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Muse Spark 1.1, left for the other.
§ 3 · Sources
Where the numbers come from
7 publications, 30 figures. Every one links to the page it was read from.
LMArena Text 1491
GDPval-AA 1217
AA-Briefcase 851
LiveBench 1.0 75.3LiveBench 1.0 · Reasoning 87.7LiveBench 1.0 · Coding 77.2LiveBench 1.0 · Agentic Coding 58.5LiveBench 1.0 · Mathematics 87.1LiveBench 1.0 · Data Analysis 72.5LiveBench 1.0 · Language 74.3LiveBench 1.0 · Instruction Following 69.6
OSWorld 80.7%
Vals · Legal Research Bench 37.98%Vals · LegalBench 84.98%Vals · Finance Agent 57.21%Vals · CorpFin 71.29%Vals · TaxEval 79.72%Vals · MortgageTax 66.14%Vals · MedScribe 88.89%Vals · SWE-bench Verified 82% (Mini-SWE-agent)Vals · Vibe Code Bench 72.16% (OpenHands)Vals · Code Migration 31.11%Vals · GPQA Diamond 91.16%Vals · MMLU Pro 88.73%
Enkrypt · Jailbreak risk 1.4%Enkrypt · Harmful content risk 0%Enkrypt · CBRN risk 6.3%Enkrypt · Toxicity risk 4.4%Enkrypt · Bias risk 50.1%Enkrypt · Insecure code risk 19.1%
Badge
[](https://publicai.io/model-index/m/muse-spark-1-1)