QQwen · Released Feb 25, 2026
Qwen3.5-27B
The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance. Its overall capabilities are comparable to those of...
- Released
- Feb 25 2026
- Context window
- 262K 66K out
- Input / 1M tokens
- $0.27 from $0.20
- Output / 1M tokens
- $2.16 from $1.56
Where to run it
6 providers compared
Cheapest: Alibaba at $0.20 / $1.56 per 1M input / output tokens — 37% less than the priciest option.
| Provider | Input | Output | Cache read | Context | Max output | Uptime | vs. best |
|---|---|---|---|---|---|---|---|
| Alibaba | $0.20 | $1.56 | — | 262K | 66K | 84.3%24h | Cheapest |
| SiliconFlowFP8 | $0.25 | $2.00 | — | 262K | 236K | 94.2%24h | +28% |
| AtlasCloudFP8 | $0.27 | $2.16 | $0.27 | 262K | 66K | 98.7%24h | +38% |
| NovitaBF16 | $0.30 | $2.40 | — | 262K | 66K | 97.8%24h | +54% |
| Phalano tool calling | $0.30 | $2.40 | $0.03 | 262K | 66K | 96.4%24h | +54% |
| DeepInfraFP8 | $0.26 | $2.60 | — | 262K | 82K | 99.8%24h | +58% |
Value
Price vs. performance
Qwen3.5-27B Other Qwen Other labs Value frontier · Vending-Bench 2 vs. input price per 1M tokens (log scale) · Drag to pan, pinch or Ctrl + scroll to zoom, double-click to zoom in.
Benchmarks
3 benchmark scores
| Benchmark | Score | Rank | Leader |
|---|---|---|---|
| Vending-Bench 2Agents & computer use · Run a simulated vending business for a year (final balance) | $202 | #54of 61 | GPT-6 Astra$15,515 |
| WebDev ArenaDesign & visual · People vote on which model builds the better website (Elo) | 1357 | #75of 88 | Claude Opus 5.51820 |
| LMCAReasoning & knowledge · Judge the quality of conceptual arguments like experts | 34.0% | #71of 96 | Claude Sonnet 5.568.3% |
Specs
Details
- Model ID
qwen/qwen3.5-27b- Developer
- Qwen
- Release date
- Feb 25, 2026 ·
- Context window
- 262,144 tokens
- Max output
- 65,536 tokens
- API price
- $0.27 / $2.16 per 1M input / output tokens
- Providers
- 6
- Input types
- text, image, video
- Reasoning
- Yes
- Tool calling
- Yes
- Open weights
- Yes
- Free tier
- No
Release date, prices and providers from OpenRouter's catalog; buying directly from a provider can cost a little differently. Uptime is the share of OpenRouter requests each provider served over the last day. Benchmark scores are the best result per benchmark from Epoch AI, licensed CC BY 4.0. Rank counts every model tracked here with a score.