23.57B Mistral scale
Captured config identifies MistralForCausalLM and Safetensors metadata reports 23572403200 parameters.
Open Source Model Profile · yentinglin
Mistral-Small-24B-Instruct-2501-reasoning
Mistral-Small-24B-Instruct-2501-reasoning is a 23.57B-parameter Mistral reasoning fine-tune from yentinglin. Its model card documents math-focused tuning on OpenR1-Math-220k and s1K-1.1 with publisher-reported MATH-500 and AIME scores.
- Publisher
- yentinglin
- Task
- text-generation
- Model type
- mistral
- License
- apache-2.0
- Library
- Unknown
- Publication status
- Accepted · not indexed
Model overview
Mistral-Small-24B-Instruct-2501-reasoning is published by yentinglin as a Mistral-based text-generation model. The captured configuration identifies MistralForCausalLM and Safetensors metadata reports 23572403200 parameters. According to the model card, it fine-tunes mistralai/Mistral-Small-24B-Instruct-2501 for mathematical reasoning, developed by Yenting Lin with Ubitus funding.
Recorded capabilities
Math-reasoning optimisation
According to the model card, the model is specifically optimised for mathematical reasoning tasks.
OpenR1-Math and s1K training
According to the model card, fine-tuning used OpenR1-Math-220k and s1K-1.1 datasets, with a config showing 5 epochs and 32768 sequence length.
Publisher-reported reasoning scores
The captured card text reports pass@1 of 0.950 on MATH-500, 0.533 on AIME 2025, and 0.667 on AIME 2024; these are publisher figures, not Ethen measurements.
H100 training hardware
According to the model card, training used 4x8 H100 GPUs provided by Ubitus.
Use cases in the source record
- Mathematical reasoning text generation evaluated with the publisher's Open-R1-based evaluation setup.
- English instruction-tuned reasoning workflows run via vLLM or sglang as described in the card.
Limitations and unknowns
- Benchmark figures are publisher-reported leaderboard values captured in the card text, not independently measured or verified by Ethen.
- No context-window value was extracted from this record as a standalone field.
- Provider state is historical snapshot data, not independently refreshed current availability.
- The publisher provides the model as-is and warns users to assess correctness themselves, excluding high-risk uses such as medical, legal, and financial advice.
Source and provenance
Source: yentinglin/Mistral-Small-24B-Instruct-2501-reasoning
Captured: Unknown. Processed: 2026-09-07T19:35:32.987669+00:00.
Mistral-Small-Reasoning This model is a fine-tuned version of mistralai/Mistral-Small-24B-Instruct-2501 , specifically optimized for mathematical reasoning tasks. It has been fine-tuned on datasets including OpenR1-Math-220k , and s1K-1.1 , aiming to enhance its reasoning capabilities. Model Details Model Description Developed by: Yenting Lin Funded by: Ubitus Model type: Instruction-tuned language model for reasoning Language(s) (NLP): English (en) License: Apache 2.0 Finetuned from model: mistralai/Mistral-Small-24B-Instruct-2501 How to Get Started with the Model A demo is available at twllm.com , and inference can be run using vL…
F001F002F003F004F005F006F007F009F010F011F012F013F014F015F016