qwen
Qwen: Qwen3.5-122B-A10B
qwen/qwen3.5-122b-a10b
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Qwen3.5-122B-A10B is a natively multimodal Qwen model that accepts text, image and video input and produces text, built on a hybrid linear-attention and sparse mixture-of-experts design with about ten billion parameters activated per token. It offers a very long native context, open weights under Apache 2.0, and compatibility with Hugging Face Transformers, vLLM and similar runtimes. Vendor-reported benchmarks place it competitively on knowledge, coding and long-context tasks, while OpenRouter's aggregates suggest weaker agentic results; independent validation evidence is thin.[1][2]
Strengths
- Natively multimodal: accepts text, image and video inputs with text output; the vendor describes early-fusion training aimed at unified vision-language capability.[1][2]
- Efficiency-oriented hybrid architecture combining Gated DeltaNet linear attention with a sparse 256-expert MoE, activating only about 10B of 122B parameters per token.[2]
- Long context of 262,144 tokens natively (extensible per the vendor to about 1,010,000), plus OpenRouter-listed support for tool calling, structured outputs and sampling controls.[1][2]
Limitations
- Performance claims rest mainly on vendor-published benchmark tables; OpenRouter's measured aggregates report a comparatively low agentic index, so results on real workloads should be validated before adoption.[1][2]
- Output is text only, reasoning is optional rather than mandatory, and no knowledge cutoff is stated in the supplied evidence.[1]
- Evidence is limited to the vendor model card and one OpenRouter catalog entry; there is no independent evaluation or third-party model card in this source pack.[1][2]
Where to use it
- Long-document or mixed-media analysis where text, image and video inputs must be processed in a single request, subject to workload testing.[1][2]
- Self-hosted deployments needing open weights, a large context window, tool calling and structured output, where the vendor reports broad language coverage.[1][2]
- Coding assistance and reasoning-oriented tasks, given vendor-reported benchmark strength, though teams should confirm quality on their own repositories.[1][2]
Sources
- [1] OpenRouter catalog · Qwen: Qwen3.5-122B-A10BSources checked: Nov 13, 2026
- [2] Model documentation · Qwen: Qwen3.5-122B-A10BSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
The Qwen3.5 122B-A10B 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. In terms of...
Specifications
- Model ID
- qwen/qwen3.5-122b-a10b
- Canonical ID
- qwen/qwen3.5-122b-a10b-20260224
- Creator
- qwen
- Added to OpenRouter
- 25.02.2026
- Context length
- 262,144
- Modality
- text+image+video->text
- Input modalities
- text, image, video
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.260 / out $2.08
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Coding | BB — next 30% of the cohort | — |
| Agents | CC — below the B band | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | input: video |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 45.7 | BB — next 30% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 7.6 | CC — below the B band | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| SiliconFlowfp8 | 262,144 | $0.260 | $2.08 | — | — | 100% |
| Alibaba | 262,144 | $0.260 | $2.08 | — | — | 100% |
| AtlasCloudfp8 | 262,144 | $0.300 | $2.4 | — | — | 100% |
| Novitabf16 | 262,144 | $0.400 | $3.2 | — | — | — |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.