QQwen · Released Feb 25, 2026
Qwen3.5-35B-A3B
The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall...
- Released
- Feb 25 2026
- Context window
- 262K 82K out
- Input / 1M tokens
- $0.14 from $0.056
- Output / 1M tokens
- $1.00 from $0.45
Where to run it
24 providers compared
Cheapest: AIHubMix at $0.056 / $0.45 per 1M input / output tokens — 77% less than the priciest option.
| Provider | Input | Output | Cache read | Context | Max output | Uptime | vs. best |
|---|---|---|---|---|---|---|---|
| AIHubMix | $0.056 | $0.45 | — | 262K | 66K | — | Cheapest |
| Merge Gateway | $0.057 | $0.46 | $0.02 | 131K | 33K | — | +2% |
| OrcaRouter | $0.057 | $0.46 | — | 262K | 66K | — | +2% |
| EmpirioLabs AI | $0.057 | $0.46 | $0.057 | 256K | 64K | — | +2% |
| 302.AI | $0.06 | $0.46 | — | 262K | 66K | — | +3% |
| SiliconFlow (China) | $0.06 | $0.48 | — | 262K | 66K | — | +6% |
| DarkbloomFP4 | $0.065 | $0.75 | $0.033 | 262K | 33K | 99.9%24h | +53% |
| DeepInfraFP8 · no tool calling | $0.14 | $1.00 | $0.05 | 262K | 82K | 99.8%24h | +129% |
| NanoGPT | $0.14 | $1.00 | $0.05 | 260K | 66K | — | +129% |
| Requesty | $0.14 | $1.00 | $0.05 | 262K | 66K | — | +129% |
| ParasailFP8 | $0.15 | $1.00 | $0.05 | 262K | 236K | 100%24h | +134% |
| Venice | $0.15 | $1.00 | $0.05 | 256K | 16K | 100%24h | +134% |
| Venice AI | $0.15 | $1.00 | $0.05 | 256K | 16K | — | +134% |
| Alibaba | $0.16 | $1.30 | — | 262K | 66K | 99.8%24h | +189% |
| Kilo Gateway | $0.16 | $1.30 | — | 262K | 82K | — | +189% |
| Mixlayer | $0.25 | $1.30 | — | 262K | 262K | — | +231% |
| AtlasCloudFP8 | $0.23 | $1.80 | $0.23 | 262K | 66K | 100%24h | +300% |
| SiliconFlowFP8 · no tool calling | $0.24 | $1.80 | $0.15 | 262K | 236K | 90%24h | +307% |
| Ofox | $0.29 | $1.83 | $0.29 | 262K | 66K | — | +336% |
| LLM Gateway | $0.25 | $2.00 | — | 262K | 66K | — | +344% |
| Jalapeno Cloud | $0.25 | $2.00 | — | 262K | 66K | — | +344% |
| Novita AI | $0.25 | $2.00 | — | 262K | 66K | — | +344% |
| DevPass (LLM Gateway) | $0.25 | $2.00 | — | 262K | 66K | — | +344% |
| Hugging Face | $0.25 | $2.00 | — | 262K | 66K | — | +344% |
Value
Price vs. performance
Qwen3.5-35B-A3B Other Qwen Other labs Value frontier · GPQA Diamond vs. input price per 1M tokens (log scale) · Drag to pan, pinch or Ctrl + scroll to zoom, double-click to zoom in.
Benchmarks
5 benchmark scores
| Benchmark | Score | Rank | Leader |
|---|---|---|---|
| SciCodeSoftware engineering · Research-coding problems from the natural sciences | 29.3% | #96of 102 | Claude Opus 5.566.9% |
| GPQA DiamondReasoning & knowledge · PhD-level biology, chemistry and physics questions | 83.5% | #63of 122 | Gemini 4 Argon96.4% |
| CritPtReasoning & knowledge · Open-ended, research-level physics problems | 0.6% | #86of 111 | GPT-5.6 Sol32.3% |
| LMCAReasoning & knowledge · Judge the quality of conceptual arguments like experts | 29.5% | #80of 123 | Claude Sonnet 5.568.3% |
| OTIS Mock AIMEMaths · Competition-style problems harder than MATH Level 5 | 70.0% | #82of 120 | Gemini 4 Argon100.0% |
Specs
Details
- Model ID
qwen/qwen3.5-35b-a3b- Developer
- Qwen
- Release date
- Feb 25, 2026 ·
- Context window
- 262,144 tokens
- Max output
- 81,920 tokens
- API price
- $0.14 / $1.00 per 1M input / output tokens
- Providers
- 24
- Input types
- text, image, video
- Reasoning
- Yes
- Tool calling
- Yes
- Open weights
- Yes
- Free tier
- No
Release date and prices from OpenRouter's catalog; providers from OpenRouter and models.dev. Buying directly from a provider can cost a little differently. Speed, latency and uptime are measured on OpenRouter only; 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.