EuroLLM-9B-Instruct (2512) vs EuroLLM-22B-Instruct (2512)
parameter count at an identical 32768-token context window (9.15B vs 22.64B)
Primerjaj
Primerjaj modele
Postavite dva do štiri modele enega ob drugega glede na polja, ki odločajo o uvedbi: kontekst, cena, regija, licenca, jeziki, gostovanje in kaj je preverjeno.
Editorially selected comparisons rather than every possible pair: 20.
EuroLLM-9B-Instruct (2512) vs EuroLLM-1.7B-Instruct
context window (32768 vs 4096 tokens)
EuroLLM-22B-Instruct (2512) vs EuroLLM-1.7B-Instruct
release date, the 2512 post-training generation against the 2024 checkpoint (2025-12-05 vs 2024-08-06)
OpenEuroLLM initial model release (planned) vs OpenEuroLLM final models (planned)
planned delivery date (31 December 2026 vs 31 January 2028)
EuroLLM-9B-Instruct (2512) vs OpenEuroLLM initial model release (planned)
weights downloadable against announced placeholder (open_weights true vs false)
EuroLLM-22B-Instruct (2512) vs OpenEuroLLM final models (planned)
licence recorded against unverified (apache-2.0 vs unknown)
Teuken 7B Instruct v0.6 vs EuroLLM-9B-Instruct (2512)
licence: CC-BY-NC-4.0, which forbids commercial serving, against Apache-2.0, which permits it
Pharia-1-LLM-7B-control-aligned vs EuroLLM-22B-Instruct (2512)
commercial channel: an on-premise agreement against publicly downloadable Apache-2.0 weights
Teuken 7B Instruct v0.6 vs Pharia-1-LLM-7B-control-aligned
EU21 average (0.57 project-reported against 0.417 in the OpenGPT-X evaluation)
Pharia-1-LLM-7B-control-aligned vs Pharia-1-Embedding-4608-control
output modality (text generation against embedding)
Devstral Small 2 (24B) vs Codestral 25.08
weights availability (downloadable Apache-2.0 checkpoint against API-only proprietary model)
Codestral 25.08 vs Qwen3-Coder-30B-A3B-Instruct
published parameter count (none published against 30.5B)
Devstral Small 2 (24B) vs GLM-4.7-Flash
published SWE-bench Verified score (68.0% against 59.2%, both vendor-reported)
Mistral Large 3 (675B) vs Mistral Small 4 (119B-A6B)
parameter count at an identical 262144-token context window (675B vs 119.4B)
Mistral Small 4 (119B-A6B) vs Ministral 3 8B
parameter count at an identical 262144-token context window (119.4B vs 8.92B)
Magistral Small 1.2 vs Devstral Small 2 (24B)
task specialisation at 24B: reasoning with explicit think tokens and vision input against agentic coding, text-only
Pixtral 12B vs Voxtral Small 24B
non-text input modality (vision against audio)
Mistral NeMo 12B Instruct vs Mistral 7B Instruct v0.3
context window (131072 vs 32768 tokens)
Mistral Small 4 (119B-A6B) vs EuroLLM-9B-Instruct (2512)
context window (262144 vs 32768 tokens)
Mistral Small 4 (119B-A6B) vs Gemma 4 31B IT
provider_slug (mistral against google), with the same Apache-2.0 licence and 262144-token context window
Kjer se dva modela razlikujeta, ta tabela pokaže razliko in se ustavi. LLM EU ne objavi značke zmagovalca brez dokumentirane ocene za to nalogo, manjkajoča številka pa ostane manjkajoča, namesto da bi postala nič.