Nous: DeepHermes 3 Mistral 24B Preview

Text input Text output
Author's Description

DeepHermes 3 (Mistral 24B Preview) is an instruction-tuned language model by Nous Research based on Mistral-Small-24B, designed for chat, function calling, and advanced multi-turn reasoning. It introduces a dual-mode system that toggles between intuitive chat responses and structured “deep reasoning” mode using special system prompts. Fine-tuned via distillation from R1, it supports structured output (JSON mode) and function call syntax for agent-based applications. DeepHermes 3 supports a **reasoning toggle via system prompt**, allowing users to switch between fast, intuitive responses and deliberate, multi-step reasoning. When activated with the following specific system instruction, the model enters a *"deep thinking"* mode—generating extended chains of thought wrapped in `<think></think>` tags before delivering a final answer. System Prompt: You are a deep thinking AI, you may use extremely long chains of thought to deeply consider the problem and deliberate with yourself via systematic reasoning processes to help come to a correct solution prior to answering. You should enclose your thoughts and internal monologue inside <think> </think> tags, and then provide your solution or response to the problem.

Key Specifications
Cost
$$
Context
32K
Parameters
24B
Released
May 09, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Reasoning Include Reasoning Seed Top P Temperature Top Logprobs Logit Bias Logprobs Stop Min P Max Tokens Frequency Penalty Presence Penalty
Features

This model supports the following features:

Reasoning
Performance Summary

Nous: DeepHermes 3 Mistral 24B Preview demonstrates a strong overall performance profile, particularly excelling in reliability and cost-effectiveness. It performs among the fastest models, ranking in the 67th percentile for speed, and offers competitive pricing, placing in the 73rd percentile. Notably, the model exhibits exceptional reliability with a perfect 100% success rate across all benchmarks, indicating consistent and stable operation. In terms of specific capabilities, DeepHermes 3 shows remarkable strength in Ethics, achieving 100% accuracy, making it the most accurate model at its price point and among models of similar speed. It also performs very well in General Knowledge (98.0% accuracy) and Email Classification (97.0% accuracy). Its Coding performance is solid at 83.0% accuracy. However, the model's Instruction Following (50.5% accuracy) and Reasoning (60.0% accuracy) capabilities are moderate, suggesting areas for potential improvement. Its hallucination rate is relatively low at 92.0% accuracy, indicating a reasonable ability to acknowledge uncertainty. The unique "deep thinking" mode, activated via a system prompt, offers a promising avenue for enhancing its reasoning capabilities, though its impact on these specific baseline scores is not directly quantifiable here.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.13
Completion $0.51

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Chutes
Chutes | nousresearch/deephermes-3-mistral-24b-preview 32K $0.13 / 1M tokens $0.51 / 1M tokens
Chutes
Chutes | nousresearch/deephermes-3-mistral-24b-preview 32K $0.13 / 1M tokens $0.51 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by nousresearch