Multilanguage multientry database design

I'm having some trouble with ecommerce product database design.

There can be dozens of languages and dozens of different fields (name, description, meta_description, meta_keywords, slug, you name it).

Does the 4th option have a huge performance hit or can it stay fast when there's hundreds of thousands rows in the translations and products tables?

1.

products_table
id, name_en, name_es, description_en, description_es, price

SELECT * FROM products_table;

2.

products_table
id, price

translations_table
id, product_id, name_en, name_es, description_en, description_es

SELECT * FROM products_table AS p
JOIN translations_table AS t ON (p.id = t.product_id);

3.

products_table
id, price

translations_table
id, product_id, lang, name, description

SELECT * FROM products_table AS p
JOIN translations_table AS t ON (p.id = t.product_id AND t.lang = 'en');

4.

products_table
id, price

translations_table
id, product_id, lang, type, value

SELECT p.*, t.value as name, t2.value as description FROM products_table AS p
JOIN translations_table AS t ON (p.id = t.product_id AND t.lang = 'en' AND type = 'name');
JOIN translations_table AS t2 ON (p.id = t.product_id AND t.lang = 'en' AND type = 'description');

5.

products_table
id, price

names_table
id, product_id, lang, value

decsriptions_table
id, product_id, lang, value

SELECT p.*, n.value, d.value FROM products_table AS p
JOIN names_table AS n ON (p.id = n.product_id AND n.lang = 'en');
JOIN descriptions_table AS d ON (p.id = d.product_id AND d.lang = 'en');