Alibaba Cloud / Qwen Team model product
Qwen3.5 397B A17B
5-397B-A17B is Qwen's flagship Mixture-of-Experts model with 397 billion total parameters and 17 billion activated parameters.
Updated Aug 12, 2026. Default version: Qwen3.5-397B-A17B
LLMBoard score66.3Qwen3.5-397B-A17B
Coverage100%37 benchmark families
Context window262.1KTokens
Official input price$0.60Alibaba API
Qwen3.5 397B A17B Specifications
Technical details for the model's default version.
- Version
- Qwen3.5-397B-A17B
- Released
- Feb 16, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 397B
- Context window
- 262.1K
- Max output
- 65.5K
- Inputs
- audio, image, text, video
- Outputs
- text
- Open weights
- Yes
- License
- Apache 2.0
Qwen3.5 397B A17B Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Qwen3.5 397B A17B Benchmark Results
Benchmark scores for Qwen3.5-397B-A17B.
| BrowseComp-zh | 70.3% | 01 | 13 | 100.0% | C | |
| NOVA-63 | 59.1% | 01 | 11 | 100.0% | C | |
| SecCodeBench | 68.3% | 01 | 1 | 100.0% | C | |
| VITA-Bench | 49.7% | 01 | 10 | 100.0% | C | |
| BFCL-V4 | 72.9% | 02 | 13 | 91.7% | C | |
| C-Eval | 93.0% | 02 | 18 | 94.1% | C | |
| LongBench v2 | 63.2% | 02 | 17 | 93.8% | C | |
| MMLU-Redux | 94.9% | 02 | 48 | 97.9% | C | |
| DeepPlanning | 34.3% | 03 | 9 | 75.0% | C | |
| HMMT25 | 92.7% | 03 | 25 | 91.7% | C | |
| Include | 85.6% | 03 | 31 | 93.3% | C | |
| MAXIFE | 88.2% | 03 | 11 | 80.0% | C | |
| MMLU-ProX | 84.7% | 03 | 32 | 93.5% | C | |
| PolyMATH | 73.3% | 04 | 23 | 86.4% | C | |
| Seal-0 | 46.9% | 04 | 6 | 40.0% | C | |
| t2-bench | 86.7% | 04 | 23 | 86.4% | C | |
| AA-LCR | 68.7% | 05 | 16 | 73.3% | C | |
| Global PIQA | 89.8% | 05 | 13 | 66.7% | C | |
| MMLU-Pro | 87.8% | 05 | 129 | 96.9% | C | |
| Multi-Challenge | 67.6% | 05 | 29 | 85.7% | C | |
| SuperGPQA | 70.4% | 05 | 34 | 87.9% | C | |
| WideSearch | 74.0% | 05 | 9 | 50.0% | C | |
| MCP-Mark | 46.1% | 06 | 8 | 28.6% | C | |
| HMMT 2025 | 94.8% | 07 | 33 | 81.3% | C | |
| IFEval | 92.6% | 08 | 65 | 89.1% | C | |
| WMT24++ | 78.9% | 08 | 23 | 68.2% | C | |
| IFBench | 76.5% | 09 | 29 | 71.4% | C | |
| AIME 2026 | 91.3% | 11 | 18 | 41.2% | C | |
| LiveCodeBench v6 | 83.6% | 12 | 53 | 78.8% | C | |
| IMO-AnswerBench | 80.9% | 14 | 19 | 27.8% | C | |
| MMMLU | 88.5% | 17 | 49 | 66.7% | C | |
| SWE-bench Multilingual | 69.3% | 20 | 34 | 42.4% | C | |
| Toolathlon | 38.3% | 26 | 31 | 16.7% | C | |
| BrowseComp | 69.0% | 29 | 58 | 50.9% | C | |
| GPQA | 88.4% | 29 | 234 | 88.0% | C | |
| SWE-Bench Verified | 76.4% | 29 | 105 | 73.1% | C | |
| Terminal-Bench 2.0 | 52.5% | 32 | 49 | 35.4% | C | |
| Humanity's Last Exam | 28.7% | 45 | 93 | 52.2% | C |
Qwen3.5 397B A17B Arena Results
Preference and agent-evaluation results for the default version.
| vision | overall | 34 | 1265.0 | 23,868 | N/A | |
| vision style control | overall | 42 | 1247.5 | 23,868 | N/A | |
| webdev | overall | 60 | 1400.3 | 17,571 | N/A | |
| text | overall | 62 | 1438.3 | 66,158 | N/A | |
| text factuality | overall | 64 | 1445.1 | 65,827 | N/A | |
| text style control | overall | 75 | 1442.0 | 66,158 | N/A |
Qwen3.5 397B A17B Runtime Performance
Provider-specific output speed and catalog latency for Qwen3.5-397B-A17B. Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Qwen3.5 397B A17B Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.60 input, $3.6 output per 1M
- Official provider
- Alibaba
- Lowest third-party
- From $0.172 input, $1.03 output per 1M via EmpirioLabs AI
- Tracked offerings
- 27
| Nvidia | qwen/qwen3.5-397b-a17b | global | N/A | N/A | 262.1K | |
| UnoRouter | qwen3.5-397b-a17b:free | global | N/A | N/A | 262.1K | |
| EmpirioLabs AI | qwen3-5-397b-a17b | global | $0.172 | $1.03 | 256K | |
| OrcaRouter | qwen/qwen3.5-397b-a17b | global | $0.172 | $1.03 | 262.1K | |
| Merge Gateway | qwen/qwen3.5-397b-a17b | global | $0.172 | $1.03 | 131.1K | |
| Vultr | Qwen/Qwen3.5-397B-A17B | global | $0.30 | $2 | 262.1K | |
| DigitalOcean | qwen3.5-397b-a17b | global | $0.3025 | $1.93 | 131.1K | |
| CrofAI | qwen3.5-397b-a17b | global | $0.35 | $1.75 | 262.1K | |
| SiliconFlow | Qwen/Qwen3.5-397B-A17B | global | $0.39 | $2.34 | 262.1K | |
| Kilo Gateway | qwen/qwen3.5-397b-a17b | global | $0.39 | $2.34 | 262.1K | |
| Alibaba (China) | qwen3.5-397b-a17b | global | $0.43 | $2.58 | 262.1K | |
| Deep Infra | Qwen/Qwen3.5-397B-A17B | global | $0.45 | $3 | 262.1K | |
| OpenRouter | qwen/qwen3.5-397b-a17b | global | $0.50 | $3.6 | 262.1K | |
| Ofox | bailian/qwen3.5-397b-a17b | global | $0.55 | $3.5 | 256K | |
| Nebius Token Factory | Qwen/Qwen3.5-397B-A17B | global | $0.60 | $3.6 | 262.1K | |
| NanoGPT | qwen/qwen3.5-397b-a17b | global | $0.60 | $3.6 | 258K | |
| LLM Gateway | qwen35-397b-a17b | global | $0.60 | $3.6 | 262.1K | |
| Mixlayer | qwen/qwen3.5-397b-a17b | global | $0.60 | $3.6 | 262.1K | |
| Scaleway | qwen3.5-397b-a17b | global | $0.60 | $3.6 | 256K | |
| Together AI | Qwen/Qwen3.5-397B-A17B | global | $0.60 | $3.6 | 262.1K | |
| Alibaba | qwen3.5-397b-a17b | global | $0.60 | $3.6 | 262.1K | |
| NovitaAI | qwen/qwen3.5-397b-a17b | global | $0.60 | $3.6 | 262.1K | |
| Hugging Face | Qwen/Qwen3.5-397B-A17B | global | $0.60 | $3.6 | 262.1K | |
| Cortecs | qwen3.5-397b-a17b | global | $0.668 | $4.01 | 262K | |
| OVHcloud AI Endpoints | qwen3.5-397b-a17b | global | $0.71 | $4.25 | 262.1K | |
| GreenPT | qwen3.5-397b-a17b | global | $0.798 | $4.96 | 262.1K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Qwen3.5 397B A17B Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Qwen3.5-397B-A17B | 66.3 | 397B | 262.1K | 65.5K | Yes | Apache 2.0 |
What is Qwen3.5 397B A17B?
Key information about Qwen3.5 397B A17B and its available data.
5-397B-A17B is Qwen's flagship Mixture-of-Experts model with 397 billion total parameters and 17 billion activated parameters. It delivers state-of-the-art performance across knowledge, reasoning, coding, mathematics, multilingual understanding, instruction following, long context, and agent tasks.
Data as of 2026-08-11.
Qwen3.5 397B A17B vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Qwen3.5 397B A17B
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
AC
Qwen3.6 Plus
Alibaba Cloud / Qwen Team
AC
Qwen3.6 27B
Alibaba Cloud / Qwen Team
AC
Qwen3.5 122B A10B
Alibaba Cloud / Qwen Team
AC
Qwen3.7 Plus
Alibaba Cloud / Qwen Team
AC
Qwen3.5 27B
Alibaba Cloud / Qwen Team
AC
Qwen3.6 35B A3B
Alibaba Cloud / Qwen Team
FAQ
Common questions about Qwen3.5 397B A17B.
When was Qwen3.5 397B A17B released?
Qwen3.5 397B A17B's default version was released on Feb 16, 2026.
How much does Qwen3.5 397B A17B cost?
Qwen3.5 397B A17B's official API price is $0.60 per million input tokens and $3.6 per million output tokens via Alibaba. The lowest tracked third-party offer starts at $0.172 input and $1.03 output via EmpirioLabs AI.
Who created Qwen3.5 397B A17B?
Qwen3.5 397B A17B was created by Alibaba Cloud / Qwen Team.
What is the context window for Qwen3.5 397B A17B?
The default version has a 262.1K token context window.
Is Qwen3.5 397B A17B open weight?
Yes. The default version is marked as open weight under Apache 2.0.
How many API providers offer Qwen3.5 397B A17B?
27 provider offerings are linked to the default version.
What models should I compare Qwen3.5 397B A17B with?
Nearby ranked alternatives include Gemini 3 Pro, GLM 5, Kimi K2.7 Code.