Is GPT-OSS Good? A Comprehensive Evaluation of OpenAI's Latest Open Source Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914199185129472 |
|---|---|
| author | Bi, Ziqian Chen, Keyu Tseng, Chiung-Yi Zhang, Danyang Wang, Tianyang Luo, Hongying Chen, Lu Huang, Junming Guan, Jibin Hao, Junfeng Song, Xinyuan Song, Junhao |
| author_facet | Bi, Ziqian Chen, Keyu Tseng, Chiung-Yi Zhang, Danyang Wang, Tianyang Luo, Hongying Chen, Lu Huang, Junming Guan, Jibin Hao, Junfeng Song, Xinyuan Song, Junhao |
| contents | In August 2025, OpenAI released GPT-OSS models, its first open weight large language models since GPT-2 in 2019, comprising two mixture of experts architectures with 120B and 20B parameters. We evaluated both variants against six contemporary open source large language models ranging from 14.7B to 235B parameters, representing both dense and sparse designs, across ten benchmarks covering general knowledge, mathematical reasoning, code generation, multilingual understanding, and conversational ability. All models were tested in unquantised form under standardised inference settings, with statistical validation using McNemars test and effect size analysis. Results show that gpt-oss-20B consistently outperforms gpt-oss-120B on several benchmarks, such as HumanEval and MMLU, despite requiring substantially less memory and energy per response. Both models demonstrate mid-tier overall performance within the current open source landscape, with relative strength in code generation and notable weaknesses in multilingual tasks. These findings provide empirical evidence that scaling in sparse architectures may not yield proportional performance gains, underscoring the need for further investigation into optimisation strategies and informing more efficient model selection for future open source deployments. More details and evaluation scripts are available at https://ai-agent-lab.github.io/gpt-oss (Project Webpage). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_12461 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Is GPT-OSS Good? A Comprehensive Evaluation of OpenAI's Latest Open Source Models Bi, Ziqian Chen, Keyu Tseng, Chiung-Yi Zhang, Danyang Wang, Tianyang Luo, Hongying Chen, Lu Huang, Junming Guan, Jibin Hao, Junfeng Song, Xinyuan Song, Junhao Computation and Language In August 2025, OpenAI released GPT-OSS models, its first open weight large language models since GPT-2 in 2019, comprising two mixture of experts architectures with 120B and 20B parameters. We evaluated both variants against six contemporary open source large language models ranging from 14.7B to 235B parameters, representing both dense and sparse designs, across ten benchmarks covering general knowledge, mathematical reasoning, code generation, multilingual understanding, and conversational ability. All models were tested in unquantised form under standardised inference settings, with statistical validation using McNemars test and effect size analysis. Results show that gpt-oss-20B consistently outperforms gpt-oss-120B on several benchmarks, such as HumanEval and MMLU, despite requiring substantially less memory and energy per response. Both models demonstrate mid-tier overall performance within the current open source landscape, with relative strength in code generation and notable weaknesses in multilingual tasks. These findings provide empirical evidence that scaling in sparse architectures may not yield proportional performance gains, underscoring the need for further investigation into optimisation strategies and informing more efficient model selection for future open source deployments. More details and evaluation scripts are available at https://ai-agent-lab.github.io/gpt-oss (Project Webpage). |
| title | Is GPT-OSS Good? A Comprehensive Evaluation of OpenAI's Latest Open Source Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2508.12461 |