X · by RamV2003
Jev against Llama and DeepSeek
235 ms median latency and 82% accuracy on the same labelled inputs.
Ram Vinjamuri
@RamV2003
(1/5) trying out jev this week with all the hype. kept it apples to apples vs the fast models it competes with: llama 3.1 8b and deepseek v4 flash same labelled inputs. jev: 235ms median, 82% accurate. llama: 497ms, 68%. deepseek flash was accurate but ~2s genuinely cool. then some odd emergent behaviour: no position or label invariance
Sep 23, 2026 · 7 likesOpen on X
RamV2003 shares early results from trying Jev: with the same labelled inputs it showed 235 ms median latency and 82% accuracy, next to Llama and DeepSeek.
Open the source
