LOADING DATASHEET
GROUNDED IN REAL HUMAN CONVERSATIONSYOUR BENCH
Qwen3.5 397B A17B
◆ OPEN
by Alibaba
OPEN WEIGHTSLOCALAPILONG CONTEXTPAIDREASONING
RANKED #189 OF 316 ASSISTANTS · OVERALL #278 OF 493 RANKED · VIBE SCORE 6.0 · 117 VOICES ·
COPY BADGE pastes a README snippet with the rank SVG.
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...
HAPPY9.8
7 mentions
SAD5.2
9 mentions
VIBE TRAJECTORY
26 weeks · 102 voices in this window
HOLDING UP
Recent vibe is steady. No clear fade in the crowd.
HOW TO USE IT
Open weights. You can use a hosted service, or download it and run it yourself, free.
RUN IT ON YOUR OWN MACHINE
Open weights: download Qwen3.5 397B A17B once and it's yours. No subscription, no rate limits, works offline.
397B total parameters (17B active MoE), a Q4 quantization requires about 240 GB to 260 GB of memory
- BARE MINIMUM
Custom PC with AMD Ryzen 9 7900X, 256 GB DDR5 RAM, and RTX 3
256 GB RAM + 12 GB VRAM
~$1,700 USEDQ2_K CPU RAM offload, ~2 tok/s, 8k context
- SWEET SPOT
Apple Mac Studio M2 Ultra (192 GB) paired with secondary Mac
192 GB UNIFIED MEMORY
~$6,500 USEDQ3_K_M unified memory, ~12 tok/s, 32k context
- FULL POWER
Workstation with 8x NVIDIA RTX 3090 24GB
192 GB VRAM
~$7,200 USEDQ4_K_M or INT4 AutoRound fully GPU-accelerated, ~28 tok/s, 64k context
LEARN HOW TO SET IT UP
WILL MY BOX RUN QWEN3.5 397B A17B?
The three cards above are researched picks. Search the machine you actually have — we only say yes if it should reply at a usable speed, not just load the weights and crawl.
HARDWARE PICKS & PRICES ARE RESEARCHED FROM THE LIVE WEB AND REFRESHED AUTOMATICALLY. TREAT THEM AS BALLPARK, NOT GOSPEL.
WHAT CAN YOU ACTUALLY DO WITH IT?
Everyday jobs on each of those rigs: how long Qwen3.5 397B A17B takes, and how good the crowd says it is at that kind of work.
| THE JOB | BARE MINIMUMCustom PC with AMD Ryzen 9 7900X, 256 GB DDR5 RAM, and RTX 3 | SWEET SPOTApple Mac Studio M2 Ultra (192 GB) paired with secondary Mac | FULL POWERWorkstation with 8x NVIDIA RTX 3090 24GB |
|---|---|---|---|
| Write an emaila paragraph or two, drafted from a one-line brief | ~4.2 MIN | ~42 S | ~18 S |
| Summarize a documenta long report boiled down to the points that matter | ~13 MIN★★★★★ | ~2.1 MIN★★★★★ | ~54 S★★★★★ |
| Build a websitea small landing page, markup and styles together | ~67 MIN★★★★★ | ~11 MIN★★★★★ | ~4.8 MIN★★★★★ |
| Build a backendan API with routes, storage and tests | ~3.5 HR★★★★★ | ~35 MIN★★★★★ | ~15 MIN★★★★★ |
| Build a gamea playable browser game in one file | ~2.1 HR★★★★★ | ~21 MIN★★★★★ | ~8.9 MIN★★★★★ |
STARS ARE THE COMMUNITY SCORE FOR THAT KIND OF WORK, DOCKED FOR HOW SQUASHED THE WEIGHTS GET AT EACH TIER. TIMES ARE COMPUTED FROM RESEARCHED SPEEDS. BALLPARK, NOT GOSPEL.
WHAT THE INTERNET SAYS
Not enough discussion about Qwen3.5 397B A17B yet to call it. We found 117 posts but people did not say much either way.
105 POSTS · 12 COMMENTS · X · GITHUB · REDDIT · DEV FORUMS · BLUESKY · HACKER NEWS · LEMMY · DEV.TO
PEOPLE LOVE
- No fan mail yet. The crowd is still warming up.
PEOPLE COMPLAIN ABOUT
- No complaints yet. The internet is still looking for one. Suspicious.
LOOKING FOR AN ALTERNATIVE?
Other models the crowd has fully reviewed, starting with multimodal models like this one.
Grok 4.6> SOURCES (12)
- REDDIT · 2026-04-22 · 598 UPVOTESQwen3.6-27B is out now!
Qwen3.6-27B is out now!
- REDDIT · 2026-02-16 · 299 UPVOTESQwen3.5 is out now!
Qwen3.5 is out now!
- REDDIT · 2026-02-16 · 228 UPVOTESQwen 3.5
Qwen 3.5
- REDDIT · 2026-02-17 · 146 UPVOTESAll Qwen3.5-397B-A17B GGUFs are up!
All Qwen3.5-397B-A17B GGUFs are up!
- REDDIT · 2026-02-20 · 103 UPVOTESQwen3.5 GGUF Evaluation Results
Qwen3.5 GGUF Evaluation Results
- REDDIT · 2026-02-16 · 94 UPVOTESQwen3.5-397B-A17B: First open-weight model in Qwen3.5 series released with benchmarks
Qwen3.5-397B-A17B: First open-weight model in Qwen3.5 series released with benchmarks
- REDDIT · 2026-03-23 · 69 UPVOTESNew LLM Debate Benchmark: models debate the same motion twice with sides swapped in 10 turns. A wide variety of controversial and relevant…
New LLM Debate Benchmark: models debate the same motion twice with sides swapped in 10 turns. A wide variety of controversial and relevant topics. Sonnet 4.6 (high) wins. GLM-5 is the open weights leader.
- REDDIT · 2026-08-02 · 51 UPVOTESAll Qwen model oneshots: 1109 outputs to look at and compare!
All Qwen model oneshots: 1109 outputs to look at and compare!
- REDDIT · 2026-02-17 · 42 UPVOTESSonnet 4.6 scores on the Extended NYT Connections benchmark
Sonnet 4.6 scores on the Extended NYT Connections benchmark
- REDDIT · 2026-03-15 · 26 UPVOTESProfessional-grade local AI on consumer hardware — 80B stable on 44GB mixed VRAM (RTX 5060 Ti ×2 + RTX 3060) for under €800 total. Full…
Professional-grade local AI on consumer hardware — 80B stable on 44GB mixed VRAM (RTX 5060 Ti ×2 + RTX 3060) for under €800 total. Full compatibility matrix included.
- REDDIT · 2026-04-09 · 24 UPVOTESWhen are we gonna get more 1-Bit models(Medium & Large size)?
When are we gonna get more 1-Bit models(Medium & Large size)?
- REDDIT · 2026-02-16 · 15 UPVOTESQwen3.5 - 397B-A17B released - open weights and available on OpenRouter
Qwen3.5 - 397B-A17B released - open weights and available on OpenRouter