Qwen3.6-35B-A3B
Released 2026-04 · hybrid · 262K tokens · 7 benchmarks
Editorial notes
Lanzado 16 abril 2026. MoE 35B total / 3B activos. Arquitectura hibrida 10x(3x(DeltaNet+MoE)→1x(Attention+MoE)). Excelente ratio quality/cost para self-hosting. Apache 2.0.
Spec sheet
- Empresa
- Alibaba
- Pais
- CN
- Tipo
- hybrid
- Release
- 2026-04
- Context
- 262K tokens
- Slug
- qwen3-6-35b-a3b
Benchmarks (7)
Coding 4
- 80.4LiveCodeBenchProblemas de coding contests en vivo de LeetCode/Codeforces.
- 73.4SWE-bench-VerifiedIssues reales de GitHub de 12 repos populares de Python.
- 51.5Terminal-Bench-2Terminal Bench v2 - tareas agenticas en CLI.
- 49.5SWE-bench-ProVersion profesional de SWE-bench con issues mas complejos.
Cite this model
BibTeX · APA
BibTeX
@misc{frontier-qwen3-6-35b-a3b,
title = {Qwen3.6-35B-A3B},
author = {{Alibaba}},
year = {2026},
note = {Frontier Benchmarks AI atlas. Accessed 2026-05-08},
url = {https://frontierbenchmarks.com/models/qwen3-6-35b-a3b}
} APA
Alibaba (2026). Qwen3.6-35B-A3B [Large language model]. Frontier Benchmarks AI. Retrieved 2026-05-08, from https://frontierbenchmarks.com/models/qwen3-6-35b-a3b
Citation refleja la pagina del atlas, no el paper original del modelo. Para el paper, ve a la seccion "Recursos" arriba.