← Countries

Country standings · France · published July 20, 2026

France on the LMArena category boards.

As of July 20, 2026, Mistral AI's Mistral Large 3 reaches rank 53 on the vision arena board, the strongest current standing for a France organization, according to LMArena.

Source: LMArena leaderboard dataset. Country attribution from the reviewed organization registry.1 organizations, 17 models on current boards. Current through July 20, 2026.

Category standings

Best France standing per LMArena board.

Each row is the highest-ranked France organization on one LMArena category board, with that board's publish date and how many top 25 slots France organizations hold.

Best France standing per LMArena category board, dated.
BoardBest rankOrganizationModelLMArena ratingTop 25 slotsPublished
Vision arena53Mistral AIMistral Large 31225.402026-07-12
Web development arena64Mistral AIMistral Large 31222.402026-05-04
Math arena73Mistral AIMistral Medium 3.51432.602026-07-20
text_style_control:instruction_following78Mistral AIMistral Medium 3.51421.102026-07-20
Coding arena79Mistral AIMistral Medium 3.51479.102026-07-20
text_style_control:multi_turn85Mistral AIMistral Medium 3.51431.802026-07-20
Overall text arena85Mistral AIMistral Medium 3.51427.602026-07-20
text_style_control:creative_writing88Mistral AIMistral Medium 3.51394.402026-07-20
text_style_control:hard_prompts91Mistral AIMistral Medium 3.51445.402026-07-20
text_style_control:longer_query96Mistral AIMistral Medium 3.51431.102026-07-20

Organizations

France organizations on the boards.

1 organizations registered to France appear on the current LMArena boards. Each row shows how many boards the organization reaches and its single best standing.

France organizations with board coverage and best standing.
OrganizationBoardsBest standing (LMArena)ModelPublished
Mistral AI10#53 Vision arenaMistral Large 32026-07-12

Frontier models

Highest-ranked France models on current boards.

The 12 best-placed models attributed to France, ordered by their best board rank, out of17 distinct France models on the current boards.

Highest-ranked France models by best board rank.
ModelOrganizationBest boardRankLMArena ratingPublished
Mistral Large 3Mistral AIVision arena#531225.42026-07-12
Mistral Medium 3.5Mistral AIVision arena#541221.82026-07-12
Mistral Medium 2508Mistral AIVision arena#771171.72026-07-12
Mistral Medium 2505Mistral AIVision arena#851157.42026-07-12
Mistral Small 2506Mistral AIVision arena#901142.32026-07-12
Mistral Small 3.1 24b Instruct 2503Mistral AIVision arena#921136.92026-07-12
Magistral Medium 2506Mistral AICoding arena#1891387.12026-07-20
Mistral Large 2407Mistral AItext_style_control:creative_writing#2201287.12026-07-20
Mistral Large 2411Mistral AItext_style_control:instruction_following#2331294.52026-07-20
Mistral Small 24b Instruct 2501Mistral AIMath arena#2531261.62026-07-20
Mistral Large 2402Mistral AIMath arena#2651244.42026-07-20
Mixtral 8x22b Instruct V0.1Mistral AIMath arena#2721228.12026-07-20

France questions

How many France organizations rank on the LMArena boards?

1 France organizations appear on the current LMArena category boards with 17 distinct models, drawn from 2 organizations registered to France. Figures are current through July 20, 2026, per LMArena.

Which France model ranks highest right now?

Mistral AI's Mistral Large 3 is the highest-ranked France model, at rank 53 on the LMArena vision arena board with a 1225.4 rating, published July 12, 2026, per LMArena.

How does Model Gauntlet decide a model belongs to France?

Model Gauntlet attributes a model to a country through its organization. Each organization carries reviewed headquarters country codes in the organization registry, and a model inherits the country codes of the organization that published it. Organizations based in more than one country are counted under each of their codes. This is organizational attribution. It is not a claim about where a model was trained or where its compute ran.

Is this a Model Gauntlet ranking of France?

No. These standings mirror the named LMArena category boards, published through July 20, 2026. Model Gauntlet runs no evaluation of its own and publishes no composite or consensus score.