Dolphin3.0 R1 Mistral 24B

Text input Text output Free Option
Author's Description

Dolphin 3.0 R1 is the next generation of the Dolphin series of instruct-tuned models. Designed to be the ultimate general purpose local model, enabling coding, math, agentic, function calling, and general use cases. The R1 version has been trained for 3 epochs to reason using 800k reasoning traces from the Dolphin-R1 dataset. Dolphin aims to be a general purpose reasoning instruct model, similar to the models behind ChatGPT, Claude, Gemini. Part of the [Dolphin 3.0 Collection](https://huggingface.co/collections/cognitivecomputations/dolphin-30-677ab47f73d7ff66743979a3) Curated and trained by [Eric Hartford](https://huggingface.co/ehartford), [Ben Gitter](https://huggingface.co/bigstorm), [BlouseJury](https://huggingface.co/BlouseJury) and [Cognitive Computations](https://huggingface.co/cognitivecomputations)

Key Specifications
Cost
$$
Context
32K
Parameters
24B
Released
Feb 13, 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

Dolphin3.0 R1 Mistral 24B, from cognitivecomputations, demonstrates exceptional speed, consistently ranking among the fastest models across seven benchmarks. It also offers highly competitive pricing, placing in the 88th percentile across six benchmarks. This instruct-tuned model, designed for general-purpose local use, aims to excel in coding, math, agentic, function calling, and general use cases. While its speed and cost-efficiency are significant advantages, the model's accuracy across various benchmarks presents a mixed picture. It struggles with hallucination detection (46.0% accuracy) and instruction following (27.2% and 0.0% accuracy in two separate tests), indicating areas for improvement in precision and adherence to complex directives. Coding performance is also relatively low at 18.2% accuracy. General knowledge (76.0% accuracy) and email classification (89.0% accuracy) show more promising results, though still ranking in the lower percentiles compared to other models. Reasoning capabilities are a notable weakness, with only 22.4% accuracy. The model's strength lies in its operational efficiency, offering rapid processing at a low cost, making it a potentially valuable tool for applications where speed and budget are paramount, provided the specific task aligns with its stronger performance areas.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.01
Completion $0.03

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Chutes
Chutes | cognitivecomputations/dolphin3.0-r1-mistral-24b 32K $0.01 / 1M tokens $0.03 / 1M tokens
Chutes
Chutes | cognitivecomputations/dolphin3.0-r1-mistral-24b 32K $0.01 / 1M tokens $0.03 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by cognitivecomputations