I gave 6 frontier LLMs the same Bach MusicXML file and prompt. The results are all one-shot and unedited
Summary
A user tested six frontier LLMs on generating music from a Bach MusicXML file in a one-shot setting, presenting unedited results.
Similar Articles
Libretto: Giving LLM Agents a Sense of Musical Structure
Libretto introduces a structured framework for symbolic music generation and revision using an LLM-native grammar and corpus-calibrated statistical evaluation across musical dimensions, enabling LLM agents to treat music as a measurable and editable object.
5 Frontier LLM's grade 39 local LLM's on 50 prompts
A hobbyist benchmarked 39 local LLMs on 50 prompts using a 5060 Ti, then had four frontier LLMs (Gemini 3.8, GPT-6.1 variants, Gemini 4) grade the outputs, ranking Bonsai-2 27B and Underdog-Saluki 27B as top in capability while naming smaller models like Nanbeige4.2 3B and Gemma-4 12B as best fine-tuning foundations.
We benchmarked 24 LLMs against human writers on 475 creative writing prompts
A benchmark release comparing 24 LLMs to human writers on 475 creative writing prompts, using a custom reward model to show that frontier models outperform amateurs but not professionals.
Evaluated 6 frontier LLMs (GPT-5.4, Claude Sonnet 4.6, Claude Opus 4.7, Gemini Pro/Flash, Grok 4.3) on political, gender, and racial bias across 8 benchmarks (~20,600 examples) [R]
A solo evaluation of six frontier LLMs on 8 bias benchmarks finds that most models lean left politically, and Grok's self-reported right-leaning stance is inconsistent with its left-leaning behavior. Refusal rates vary, with GPT-5.4 refusing 20% of race-related questions.
Tried LFM2.5-24-A2B. It's crap. Yes. I double confirm it. I wouldn't trust it peeling a peanut correctly
A LocalLLaMA user reports negative hands-on experiences with the quantized LFM2.5-24B-A2B (Q4_K_M GGUF) model, citing looping behavior, broken code generation on agentic tasks, and poor plugin development results despite using a corrected chat template and Hermes integration.