5 ms·
These tests mean nothing; I yet to see a model that is better than Sonnet 4 for coding. I tried many, all of them are sub-par, even with a small code base.
by Oras 1y ago
These tests mean nothing; I yet to see a model that is better than Sonnet 4 for coding. I tried many, all of them are sub-par, even with a small code base.
- nnevatie 1y agoWell, Codex with GPT5 High wins Claude Sonnet 4.5 - this is anecdotal, but I've used both extensively.
- solarkraft 1y agoAt what speed? At some point you’ll have to compare to Opus.
- adastra22 1y agoAnd Sonnet 4.5 is better than Opus.
- Bolwin 1y agoWell yeah no surprise. You should try glm 4.6
- Oras 1y agoI tried it, and it was shockingly bad compared to their benchmarks and to Claude Sonnet 4. I tried it with Claude Code CLI, it didn't follow instructions correctly (I had a Claude.md file with clear instructions), stopped after a few implementations (less than 3 minutes), and produced code that does not work. For the benefit of the doubt, I changed instructions to be NextJS platform as I thought it's a known framework and it might do better, but still, same quality issues.
- adastra22 1y agoWell, Sonnet 4.5 is better.