Benchmarks

Een geverifieerde claim op LLM EU zegt waar die vandaan komt: een document dat je kunt openen, of een waarde die op de endpointrij is vastgelegd.

European languages

The EU21 numbers are OpenGPT-X’s, taken from the Teuken 7B Instruct v0.6 model card and the accompanying evaluation preprint; they are machine-translated benchmarks, not native-language test sets. The MMMLU number is Google’s own. LLM EU has run no evaluation and has not verified either harness.

Modellen	score		source
Teuken 7B Instruct v0.6	0.57		link
Llama 3.1 8B Instruct	0.563		link
Mistral 7B Instruct v0.3	0.527		link
Aya 23 8B	0.485		link
Pharia-1-LLM-7B-control-aligned	0.417		link
Gemma 4 31B IT	88.4		link

Code and software agents

Every score in this set is republished from the model publisher’s own model card or vendor report and names that source per entry. LLM EU has run no evaluation, has not reproduced any score, and has not verified the harness settings. Vendor-reported numbers are self-reported and are labelled as such on the model cards. SWE-bench results depend heavily on the agent scaffold; the NVIDIA entry names its scaffold explicitly, and scores from different scaffolds must not be compared directly.

Modellen	score		source
Devstral Small 2 (24B)	68		link
GLM-4.7-Flash	59.2		link
NVIDIA Nemotron 3 Super 120B-A12B	60.47		link
Gemma 4 31B IT	80		link
Magistral Small 1.2	70.88		link
GLM-5.3	88.8		link

Reasoning and mathematics

Every score in this set is republished from the model publisher’s own model card or vendor report and names that source per entry. LLM EU has run no evaluation, has not reproduced any score, and has not verified the harness settings. Vendor-reported numbers are self-reported and are labelled as such on the model cards. AIME and GPQA scores are sampling-sensitive: number of samples, temperature and tool access are stated on the individual model cards and are not uniform across these rows.

Modellen	score		source
GLM-4.7-Flash	91.6		link
NVIDIA Nemotron 3 Super 120B-A12B	90.21		link
Gemma 4 31B IT	89.2		link
Magistral Small 1.2	77.34		link
DeepSeek-R1	79.8		link
DeepSeek-R1	71.5		link
Phi-4	56.1		link

Long context

Every score in this set is republished from the model publisher’s own model card or vendor report and names that source per entry. LLM EU has run no evaluation, has not reproduced any score, and has not verified the harness settings. Vendor-reported numbers are self-reported and are labelled as such on the model cards. RULER scores come from NVIDIA’s own NeMo Evaluator harness and are not comparable with scores produced by other harnesses. The context-window rows are configuration or vendor claims and say nothing about answer quality at that length.

Modellen	score		source
NVIDIA Nemotron 3 Super 120B-A12B	96.3		link
NVIDIA Nemotron 3 Super 120B-A12B	91.75		link
NVIDIA Nemotron 3 Nano 4B	91.1		link
Llama 4 Scout 17B-16E Instruct	10485760		link
GLM-5.3	1048576		link
DeepSeek-V4-Flash (0731)	1048576		link

Small models, cost bracket

Every score in this set is republished from the model publisher’s own model card or vendor report and names that source per entry. LLM EU has run no evaluation, has not reproduced any score, and has not verified the harness settings. Vendor-reported numbers are self-reported and are labelled as such on the model cards. LLM EU publishes no price data in this set and has run no cost or throughput benchmark, so the rows must not be read as value for money: each metric is stated per row and the models differ in size, licence and serving requirements.

Modellen	score		source
Phi-4	84.8		link
Gemma 4 E4B IT	69.4		link
Mistral NeMo 12B Instruct	68		link
Ministral 3 8B	0.668		link
NVIDIA Nemotron 3 Nano 4B	78.5		link
GLM-4.7-Flash	59.2		link

LLM EU Verified betekent gedocumenteerd, niet wettelijk gecertificeerd. Het is geen conformiteitsbeoordeling onder de AI-verordening en LLM EU is geen EU-orgaan.