ℏεsam(@Hesamation):we need to talk about how GPT-5.6 Luna beats Opus 4.8 at agentic coding, with a fraction of the cost. and it's not even far off from Fable 5. now this is on benchmarks, so take that with a pinch of salt. but the majority of coding tasks aren't much harder than bench-level.

2026.07.10 15:25

we need to talk about how GPT-5.6 Luna beats Opus 4.8 at agentic coding, with a fraction of the cost. and it's not even far off from Fable 5. now this is on benchmarks, so take that with a pinch of salt. but the majority of coding tasks aren't much harder than bench-level.

Sebastian Raschka@rasbt

2026.07.10 13:32

For agentic coding, one can say: - Unless you need Terra Ultra perf, it's always better to use a Luna model with higher effort setting (same or better performance but cheaper). - Forget everything below Sol High, use Luna with higher effort settings here - Forget Sol Extra High, use Terra Ultra here - The extra cost of Sol Ultra is probably not worth it over Max