qwen
Qwen: Qwen3 14B
qwen/qwen3-14b
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Qwen3 14B is a dense, post-trained causal language model in the Qwen3 family, distributed under the Apache-2.0 license. Its model card describes switchable thinking and non-thinking modes, multilingual instruction following and translation across more than 100 languages and dialects, and tool-calling support for agent workflows. OpenRouter lists it as text-only, with reasoning, tools and structured-output parameters exposed. These are capability claims and API surface details, not proof of performance for a specific deployment.[1][2]
Strengths
- The model card reports a hard switch between thinking and non-thinking modes; when thinking is enabled, /think and /no_think can steer behavior turn by turn.[2]
- The documentation reports support for more than 100 languages and dialects, including multilingual instruction following and translation.[2]
- The documentation describes tool calling and agent integration in both modes, while OpenRouter exposes tools, tool choice, structured outputs and reasoning parameters.[1][2]
Limitations
- The card specifies a 32,768-token native context; reaching 131,072 requires YaRN configuration, and static YaRN may reduce shorter-text performance. OpenRouter lists 40,960, so confirm the deployed endpoint.[1][2]
- Reasoning requires configuration care: the card warns against greedy decoding in thinking mode, recommends different sampling settings by mode, and says prior thinking content should not remain in conversation history.[2]
- Evidence is limited to the vendor model card and an OpenRouter catalog entry. OpenRouter reports coding and agentic indices without enough context to interpret them, so validate accuracy, latency, context behavior and tool reliability on the intended workload.[1][2]
Where to use it
- Potentially suitable for assistants that alternate routine dialogue with deliberate reasoning, including math or code help, after testing mode control.[2]
- Potentially suitable for multilingual drafting, translation and instruction-following products, subject to validation in each target language.[2]
- Potentially suitable for tool-using agents or structured-output services, after validating tool-calling reliability and endpoint parameter support.[1][2]
Sources
- [1] OpenRouter catalog · Qwen: Qwen3 14BSources checked: Nov 13, 2026
- [2] Model documentation · Qwen: Qwen3 14BSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Qwen3-14B is a dense 14.8B parameter causal language model from the Qwen3 series, designed for both complex reasoning and efficient dialogue. It supports seamless switching between a "thinking" mode for...
Specifications
- Model ID
- qwen/qwen3-14b
- Canonical ID
- qwen/qwen3-14b-04-28
- Creator
- qwen
- Added to OpenRouter
- 28.04.2025
- Context length
- 40,960
- Modality
- text->text
- Input modalities
- text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.120 / out $0.240
- Knowledge cutoff
- 31.03.2025
- 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 | CC — below the B band | — |
| 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 | — |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 13.8 | CC — below the B band | 13.11.2026 |
| Artificial Analysis | Agentic index | 0.9 | 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) |
|---|---|---|---|---|---|---|
| NextBitint4 | 40,960 | $0.100 | $0.220 | — | — | 100% |
| DeepInfrafp8 | 40,960 | $0.120 | $0.240 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.