YFarmXIntelligence for an accelerating worldBriefings
YFarmX
YFarmX

AI Desk · Model IdentityWithdrawn

Nous: Hermes 4 70B

nousresearch/hermes-4-70b · Nous ResearchDeclared and measured agree
Tokenizer lineage (measured)
Llama3
Evidence confidence
Very high
Why very high: an exact fingerprint shared with 4 models and an identical API contract inside the same group: two signals of different kinds agree.
Declared tokenizer tag
Llama3
Entered catalogue
26 August 2025
Last tested
21 August 2026
Measurement
Measured
Evidence
  • ✓ exact tokenizer signature shared with 4 models
  • ✓ API-surface contract identical to a group member
Compare with Meta: Llama 3.2 1B Instruct

Download JSONReport a correction

Withdrawn from the catalogue

First absent from the snapshot of 16 September 2026

Nous: Hermes 4 70B no longer appears in the OpenRouter catalogue. The record stays here because the measurements below were real when they were taken and this address has been published. A model leaving is the kind of change this product exists to notice, so it is kept rather than deleted.

Identity statement

Nous: Hermes 4 70B is Nous Research's model. Its measured tokenizer sits with the Llama3 group, which is evidence that it builds on the Llama3 tokenizer. Tokenizer reuse is normal engineering, open vocabularies travel between labs, and this says nothing against Nous Research's authorship of the model itself.

Evidence stack

Signals of different kinds, weighed together
SignalResultReference
Tokenizer signatureExact match with 4 other modelstk_47615f2d
API surfaceIdentical contract to 1 other entryapi_9f8a0c50
Context and output131,072 context · 117,964 max outputdeclared
Reasoning contractOptionaldeclared
Serving providers observedNebiusmeasured 21 August 2026

Shares this fingerprint

Exact signature first; template-boundary shifts of the same signature beneath
Meta: Llama 3.2 1B Instructmeta-llama/llama-3.2-1b-instructexact · 50/50Nous: Hermes 3 405B Instructnousresearch/hermes-3-llama-3.1-405bexact · 50/50Nous: Hermes 3 70B Instructnousresearch/hermes-3-llama-3.1-70bexact · 50/50Nous: Hermes 4 405Bnousresearch/hermes-4-405bexact · 50/50Meta: Llama 3.1 70B Instructmeta-llama/llama-3.1-70b-instructsame tokenizer · template shiftMeta: Llama 3.1 8B Instructmeta-llama/llama-3.1-8b-instructsame tokenizer · template shiftMeta: Llama 3.2 3B Instructmeta-llama/llama-3.2-3b-instructsame tokenizer · template shiftMeta: Llama 3.3 70B Instructmeta-llama/llama-3.3-70b-instructsame tokenizer · template shiftPerplexity: Sonarperplexity/sonarsame tokenizer · template shiftPerplexity: Sonar Properplexity/sonar-prosame tokenizer · template shiftSao10K: Llama 3.1 Euryale 70B v2.2sao10k/l3.1-euryale-70bsame tokenizer · template shiftSao10K: Llama 3.3 Euryale 70Bsao10k/l3.3-euryale-70bsame tokenizer · template shiftSao10K: Llama 3 8B Lunarissao10k/l3-lunaris-8bsame tokenizer · template shift

Closest measured models

Agreement across the mutually clean test strings
Meta: Llama 3.2 1B Instructmeta-llama/llama-3.2-1b-instruct50/50 · 100%Nous: Hermes 3 70B Instructnousresearch/hermes-3-llama-3.1-70b50/50 · 100%Nous: Hermes 3 405B Instructnousresearch/hermes-3-llama-3.1-405b50/50 · 100%Nous: Hermes 4 405Bnousresearch/hermes-4-405b50/50 · 100%Meta: Llama 3.2 3B Instructmeta-llama/llama-3.2-3b-instruct48/50 · 96%Meta: Llama 3.1 8B Instructmeta-llama/llama-3.1-8b-instruct48/50 · 96%

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-long16
spaces-203
spaces-603
tabs-203
newlines-204
mixed-ws5

Digits

digits-93
digits-124
digits-3010
digits-sep7
float-long9

CJK

zh-common12
zh-long18
zh-rare18
ja-kana11
ja-kanji12
ko14

Other scripts

ru14
ar12
he24
hi15
th14
el13

Emoji

emoji-basic10
emoji-skin30
emoji-zwj-family15
emoji-zwj-x345
emoji-flags24
emoji-prof26

Rare Unicode

math31
boxdraw23
combining11
cjk-ext-b13
surrogates40
zalgo35
rtl-mix8

Code

py-code25
py-indent15
json20
html16
regex44
camel5
snake6

Repetition and encodings

rare-word-x531
repeat-tok22
base6435
hex11
url17
uuid27

An x marks a row excluded as corrupt (caching or provider interference detected during measurement). Overhead subtracted: 24 prompt tokens.

Declared record

What the catalogue claims about this model

Hermes 4 70B is a hybrid reasoning model from Nous Research, built on Meta-Llama-3.1-70B. It introduces the same hybrid mode as the larger 405B release, allowing the model to either...

Modalities
text->text
Declared tokenizer
Llama3
Prompt price
$0.13 / M tokens
Completion price
$0.40 / M tokens

Supported parameters

frequency_penaltyinclude_reasoningmax_tokenspresence_penaltyreasoningrepetition_penaltyresponse_formattemperaturetop_ktop_p

defaults: {}

Catalogue entryWeights on Hugging Face

History

Every observation, kept as taken
  • 26 August 2025Enters the OpenRouter catalogue declared as Llama3.
  • 21 August 2026Fingerprinted in the YFarmX catalogue sweep · 50 of 50 strings measured clean.
  • 21 August 2026YFarmX assessment: consistent with the Llama3 family, very high confidence.

Cite this page as the evidence record for Nous: Hermes 4 70B: the URL is stable, measurements are dated, and revisions append to the history above. Method: tokenizer fingerprinting, v1, 50 strings.