RLHF from Qwen 2.5 72B-Instruct
According to the model card, the model was trained through RLHF with Qwen-2.5-72B-Instruct as base model and is described as a chat model.
Open Source Model Profile · Nexusflow
Athene-V2-Chat
Athene-V2-Chat is a 72.71B-parameter Qwen2-family chat model from Nexusflow. According to the model card, it is an RLHF finetune of Qwen 2.5 72B-Instruct.
- Publisher
- Nexusflow
- Task
- text-generation
- Model type
- qwen2
- License
- other
- Library
- transformers
- Publication status
- Approved for indexing
Model overview
Athene-V2-Chat is published by Nexusflow as a text-generation chat model. Safetensors metadata reports 72,706,203,648 parameters, with Qwen2ForCausalLM and model type qwen2. According to the model card, it was trained through RLHF from Qwen 2.5 72B-Instruct.
Recorded capabilities
Qwen2.5 chat template and usage
The model card says Athene-V2-Chat uses the same chat template as Qwen2.5-72B-Instruct and documents Transformers usage with a step-by-step system-prompt note.
Publisher benchmark description
The model card describes the 72B release as on-par with GPT-4o across benchmarks and discusses Chatbot Arena hard, math, coding, and instruction-following categories.
Use cases in the source record
- Conversational text-generation workflows using the documented Qwen2.5-72B-Instruct-compatible chat template.
- Chat, math, and coding experiments of the kind described in the publisher model card and Transformers example.
Limitations and unknowns
- No context-window value was extracted from this record.
- Provider state is historical snapshot data, not independently refreshed current availability.
- Benchmark and capability comparisons are publisher claims and were not independently verified by Ethen.
Source and provenance
Source: Nexusflow/Athene-V2-Chat
Captured: Unknown. Processed: 2026-09-07T19:34:34.838139+00:00.
Athene-V2-Chat-72B: Rivaling GPT-4o across Benchmarks Nexusflow HF - Nexusflow Discord - Athene-V2 Blogpost We introduce Athene-V2-Chat-72B, an open-weights LLM on-par with GPT-4o across benchmarks. It is currently the best open model according to Chatbot Arena , where it beats GPT-4o-0513 (the best GPT-4o model on Arena) in hard and math category, and is on-par with GPT-4o-0513 in coding, instruction following, longer query and multi-turn. It is trained through RLHF with Qwen-2.5-72B-Instruct as base model. Athene-V2-Chat-72B excels in chat, math, and coding. Its sister model, Athene-V2-Agent-72B , surpasses GPT-4o in complex funct…
F001F002F003F004F005F006F007F010F011F012F013F015F016F017F018