‹ PublicAI Index
The LLM benchmark aggregator.
Ling 3.0 Flash
InclusionAI · 128B · open weights
Strongest in Science (#51 of 122), weakest in Safety (#262 of 329). Above par in 7 of 18 scopes. Among the models it meets almost everywhere, it finishes behind Claude Sonnet 5 and Claude Opus 4.7 and ahead of GPT-5 nano and Grok 4 Fast.
§ 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.
Reasoning52.9−12.5#67/1951/5
Science56.1−7.4#51/1221/1
Knowledge50.4−12.1#107/1971/3
Academic knowledge51−12.9#76/1231/1
Professional47.3−15.1#131/1941/2
Legal51.1−14.8#83/1731/1
Finance46.4−16#115/1621/1
Medical45−21.5#119/1491/1
Agents49.8−16.6#131/2982/6
Knowledge work49.6−21.1#97/2082/2
Coding41.9−24.3#157/1791/6
Agentic coding41.3−25.6#153/1731/5
Safety46.3−15.1#262/3291/3
Toxicity avoidance55.8−3.2#84/2681/1
Secure code56.4−9.8#102/2701/1
Fairness43.4−26.6#230/2921/2
Harm refusal45.5−16#233/2921/2
Jailbreak resistance31.2−33.8#249/2681/1
§ 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 Ling 3.0 Flash, left for the other.
§ 3 · Sources
Where the numbers come from
4 publications, 18 figures. Every one links to the page it was read from.
GDPval-AA 957
AA-Briefcase 798
Vals · LegalBench 79.69%Vals · Finance Agent 30.31%Vals · CorpFin 61.93%Vals · TaxEval 70.65%Vals · MedCode 32.27%Vals · MedScribe 80.9%Vals · SWE-bench Verified 65.2% (Mini-SWE-agent)Vals · Vibe Code Bench 2.91% (OpenHands)Vals · GPQA Diamond 84.85%Vals · MMLU Pro 82.01%
Enkrypt · Jailbreak risk 28.6%Enkrypt · Harmful content risk 6.1%Enkrypt · CBRN risk 52.7%Enkrypt · Toxicity risk 2.2%Enkrypt · Bias risk 88.1%Enkrypt · Insecure code risk 20%
Badge
[](https://publicai.io/model-index/m/ling-3-0-flash)