LLM Digest

Story

arxiv_cs_ai · Oct 21, 2026 · paper

Source brief

SWE-DIVE: Multi-Turn Coding Agent Evaluation with Dynamic User Simulation

arxiv.orgOct 21, 2026
original source linked

In brief

Coding agents are typically evaluated using one-shot, well-formulated prompts. Benchmarks such as SWE-bench use existing GitHub issues as input for a coding agent and evaluate whether the agent successfully solves the...

Feed lens

agentevaluation

Read the original at arxiv.org →Open in live feed Read that day’s brief

SWE-DIVE: Multi-Turn Coding Agent Evaluation with Dynamic User Simulation

Earlier in this thread 4 items

GPUPhysBench: Benchmarking Coding Agents for Correct and Efficient GPU Physics Simulation

Agent Evaluation Metric for multi-turn conversations

KC-Bench: A Dynamic Interactive Benchmark for Evaluating Knowledge Conflicts in LLM Agents

Multi-turn Conversational AI from Text to Multimodal Interaction: Data, Models, Evaluation, and Open Challenges