Mistral: Mistral Medium 3

mistralai/mistral-medium-3 · Mistral AIDeclared and measured agree

Tokenizer lineage (measured): Mistral
Evidence confidence: Very high
Declared tokenizer tag: Mistral
Entered catalogue: 7 May 2025
Last tested: 22 August 2026
Measurement: Measured

Evidence

✓ exact tokenizer signature shared with 21 models
✓ API-surface contract identical to a group member

Compare with Mistral: Mistral Medium 3.5

Download JSON Report a correction

Identity statement

Mistral: Mistral Medium 3 is Mistral AI's model. Its measured tokenizer sits with the Mistral group, which is evidence that it builds on the Mistral tokenizer. Tokenizer reuse is normal engineering, open vocabularies travel between labs, and this says nothing against Mistral AI's authorship of the model itself.

Evidence stack

Signals of different kinds, weighed together

Signal	Result	Reference
Tokenizer signature	Exact match with 21 other models	tk_0da8a999
API surface	Identical contract to 3 other entries	api_f2644e32
Context and output	131,072 context · 104,857 max output	declared
Reasoning contract	None declared	declared
Serving providers observed	Mistral	measured 22 August 2026

Shares this fingerprint

Exact signature first; template-boundary shifts of the same signature beneath

Closest measured models

Agreement across the mutually clean test strings

Mistral: Mistral Nemomistralai/mistral-nemo50/50 · 100%NVIDIA: Nemotron 3 Nano 30B A3Bnvidia/nemotron-3-nano-30b-a3b50/50 · 100%NVIDIA: Nemotron 3.5 Lightningnvidia/nemotron-3.5-lightning50/50 · 100%NVIDIA: Nemotron 3 Supernvidia/nemotron-3-super-120b-a12b50/50 · 100%Mistral: Mistral Small 3.2 24Bmistralai/mistral-small-3.2-24b-instruct50/50 · 100%Mistral: Ministral 3 3B 2512mistralai/ministral-3b-251250/50 · 100%

Click a row to open the full comparison.

The measured fingerprint

Marginal prompt-token cost of each test string, grouped by script

All 50 rows

English and whitespace

en-prose14

en-long18

spaces-203

spaces-603

tabs-204

newlines-207

mixed-ws7

Digits

digits-99

digits-1212

digits-3030

digits-sep12

float-long22

CJK

zh-common16

zh-long20

zh-rare20

ja-kana11

ja-kanji13

ko13

Other scripts

ru12

ar9

he10

hi14

th17

el13

Emoji

emoji-basic20

emoji-skin40

emoji-zwj-family19

emoji-zwj-x357

emoji-flags32

emoji-prof35

Rare Unicode

math33

boxdraw30

combining14

cjk-ext-b16

surrogates53

zalgo36

rtl-mix6

Code

py-code25

py-indent15

json25

html16

regex49

camel6

snake7

Repetition and encodings

rare-word-x536

repeat-tok41

base6439

hex19

url19

uuid36

An x marks a row excluded as corrupt (caching or provider interference detected during measurement). Overhead subtracted: 3 prompt tokens.

Declared record

What the catalogue claims about this model

Mistral Medium 3 is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost. It balances state-of-the-art reasoning and multimodal performance with 8× lower cost...

Modalities: text+image+file->text
Declared tokenizer: Mistral
Prompt price: $0.40 / M tokens
Completion price: $2.00 / M tokens

Supported parameters

frequency_penaltymax_tokenspresence_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_p

defaults: {"temperature":0.3}

Catalogue entry

History

Every observation, kept as taken

7 May 2025Enters the OpenRouter catalogue declared as Mistral.
22 August 2026Fingerprinted in the YFarmX catalogue sweep · 50 of 50 strings measured clean.
22 August 2026YFarmX assessment: consistent with the Mistral family, very high confidence.

Cite this page as the evidence record for Mistral: Mistral Medium 3: the URL is stable, measurements are dated, and revisions append to the history above. Method: tokenizer fingerprinting, v1, 50 strings.