Morph
Official sourceBenchmark
Morph published a scored leaderboard of AI coding agents on morphllm.com, updated August 2, 2026, benchmarking agents on SWE-bench Verified, SWE-bench Pro, and Terminal-Bench v2 with per-task pricing. Key results: GPT-5.6 Sol at xhigh effort leads Terminal-Bench 2.1 at 89.5%, Claude Opus 5 at max effort follows at 89.1 The article frames Morph as an independent benchmark curator and explicitly states it contains no paid placements, making it a first-party reference point for developers comparing coding-agent models and total cost of ownership. Terminal-Bench 2.1 scores reflect agent-plus-model entries where tbench.ai publishes them,