Which local model should a coding agent run?

Automated excerpt

One harness, same conditions, four dimensions, and the raw receipts of every call — this round, seven GGUFs under llama.cpp plus one model under the Strata engine, each in both thinking modes. The current pool runs set v5: 18 tool tasks, 72 code tasks across four languages (32 easy, 20 hard, 20 expert), 72 multi-step agent tasks (13 of them multi-file projects) and a 32K output budget for code. It did not: 0.920 against 0.923 on the same 72 agent tasks, same model, same quant, at the same throughput (48 tok/s against 47).

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.