PipelineScore
← Back to leaderboard
nous

Hermes 3 Llama 3.1 405B

Released 2024-08-16Context 128Khermes-3-llama-3-1-405b
PipelineScore
84.0MAINLINE
Ranked #8 of 52 models · 85th percentileAn even spread: no standout, no liability (82 to 85 across all five categories). Best-fit profile: Coding.

Category breakdown

Score per category, normalized 0–100 against the v1 anchor.

Code
85.0
Reason
82.3
Tool Use
85.2
RAG
82.5
Speed
83.6

Strengths

Tool Use85.2
Code85.0
Speed83.6

Same model, different rigs

Every submission of Hermes 3 Llama 3.1 405B on the 0–100 scale. The spread is the point: where it runs changes what you get.

0255075100

Best 84.9 on b200-192gb · lowest 81.5 on a100-80gb · spread 3.4 pts across 4 runs. Hover a dot for its rig.

Sample tasks

A taste of what the test pack measures. Full prompts are private and rotated daily.

CodeDifficulty 1code-fib-1

Fibonacci function

Write a Python `fib(n)` returning the nth Fibonacci number, O(n).

ReasonDifficulty 1reason-math-1

Train meeting time

Two trains, opposite directions, given speeds and start times — when do they meet?

RAGDifficulty 2rag-extract-1

Extract metrics to JSON

From the context, extract net sales, operating margin, and free cash flow as a JSON object. Numbers only.

Tool UseDifficulty 2tool-schema-1

OpenAPI param selection

Given an OpenAPI schema with limit/offset/sort, fill JSON for 'next 50, recent first.'

RAGDifficulty 2rag-grounding-1

Refuses to fabricate

Context lacks the answer — does the model fabricate or correctly say it can't?

Compare with

Hermes 3 Llama 3.1 405B vs openai/gpt-oss-20bHermes 3 Llama 3.1 405B vs DeepSeek R1 671B-A37BHermes 3 Llama 3.1 405B vs Qwen 3 235B-A22B MoEHermes 3 Llama 3.1 405B vs DeepSeek V3 671B-A37BHermes 3 Llama 3.1 405B vs DeepSeek Coder V2 236B