Symflower

Benchmarking LLM agents 💡 Our latest post covers useful benchmarks for evaluating LLM code generation agents and agentic software development workflows.

https://symflower.com/en/company/blog/2025/benchmarks-llm-agents/

Benchmarks evaluating LLM agents for software development Benchmarks for evaluating LLM code generation agents and agentic software development workflows.

New to LLM coding agents? 🤖 Our introduction covers the capabilities, limitations, and use cases of LLM agents for software development 👇

An introduction to LLM agents for software development An introduction to LLM agents for software development and their capabilities, their limitations, and use cases.

23 is out! 🥳 Learn about all the updates & new features in :https://symflower.com/en/company/blog/2024/what-to-expect-java-23/

What is new in Java 23? Read on for descriptions of the 12 features in Java 23.

Execute only the tests you need 💡We see a 29% reduction in test ex*****on times with just a basic approach.

Details of the benchmark, example to try out yourself, and guide on how to integrate the new `symflower test-runner` command in your Go repository:
https://symflower.com/en/company/blog/2024/test-impact-analysis/

Let us know how much impact we have on your repository, and which programming languages and tools we should implement next!

Test impact analysis: Automatically run affected tests only Showcasing how even a basic test impact analysis can already save 29% test ex*****on time on average.

Need to cut costs? 🤑 Read up on the key practices you can use to optimize your LLM spending 👌

https://symflower.com/en/company/blog/2024/managing-llm-costs/

Managing the costs of LLMs An overview of the techniques to control costs when using Large Language Models to generate text or software code.

👀 Monitoring can help improve the performance of your LLM applications. Here’s what you need to know & the most useful tools for LLM observability 🔍
https://symflower.com/en/company/blog/2024/llm-observability/

symflower.com

We used to and code to 🦾 Here‘s what we experienced:

https://symflower.com/en/company/blog/2024/transpilation-with-llms/

Using LLMs to transpile code (from Go & Java to Ruby) To extend DevQualityEval with Ruby, we used LLMs to transpile Java and Go test repositories to Ruby.

Are you using -powered tools in your workflow❓ Aider is a good example that works well and even offers voice coding 🦾 Here’s our guide to using Aider:

https://symflower.com/en/company/blog/2024/using-aider-for-code-generation/

Using Aider AI for code generation Learn more about Aider, an LLM-powered coding assistant, and find out how to use it in your workflow.

Lost in the sea of tools? 🌊 We’ve got you! Here’s our list of the top tools for :

https://symflower.com/en/company/blog/2024/ai-tools-software-testing/

AI tools for software development This post covers the various LLM-based tools that support software development, with a description of key features and supported technologies.

How well do generate code ❓ There’s only one way to find out: models for tasks. Here’s a roundup of popular LLM benchmarks & insights into our take on the topic 🤓

https://symflower.com/en/company/blog/2024/comparing-llm-benchmarks/

Comparing LLM benchmarks for software development A comparison of the most widely used LLM benchmarks for coding, and insights to a new code generation benchmark.

Looking to evaluate LLMs? 👀 This post helps you navigate the landscape 🧭
https://symflower.com/en/company/blog/2024/llm-benchmarks/

What are the most popular LLM benchmarks? A comprehensive description of the most widely used LLM benchmarks.

What metrics do you track when evaluating ? 👀 Here‘s an overview of complex statistical and model-based scorers 💡 Bonus: we also cover the that help you get started assessing .

https://symflower.com/en/company/blog/2024/llm-complex-scorers-evaluation-frameworks/

Evaluating LLMs: complex scorers and evaluation frameworks An overview of unsupervised LLM evaluation metrics and the most popular frameworks to help you evaluate models for your use case.