qwen3.5-9b API — compare providers
The same model (qwen3.5-9b) is served by 5 providers; SiliconFlow is the cheapest at $0.25/1M blended — the priciest (Together AI) costs 68% more.
| Rank | Provider | Input /1M | Output /1M | Blended /1M |
|---|---|---|---|---|
| 1 | SiliconFlow | $0.1 | $0.15 | $0.25 ← cheapest |
| 2 | DeepInfra | $0.1 | $0.15 | $0.25 |
| 3 | Venice | $0.1 | $0.15 | $0.25 |
| 4 | Parasail | $0.1 | $0.25 | $0.35 |
| 5 | Together AI | $0.17 | $0.25 | $0.42 |
Prices are per 1M tokens (USD). "Blended" = input + output, for coarse ranking. Same underlying model, 5 serving providers, 68% spread top to bottom. Context window: 262,144 tokens (235,929 max output) — see how it ranks in biggest context windows. See how this compares across the catalogue in same model, different price.
Cost per 1,000 requests by workload
What each provider actually bills for a representative job, not just the sticker price.
| Provider | Chatbot 1000 in / 500 out | RAG / long context 8000 in / 500 out | Batch summarize 4000 in / 1000 out |
|---|---|---|---|
| SiliconFlow | $0.17 | $0.88 | $0.55 |
| DeepInfra | $0.17 | $0.88 | $0.55 |
| Venice | $0.17 | $0.88 | $0.55 |
| Parasail | $0.23 | $0.93 | $0.65 |
| Together AI | $0.30 | $1.49 | $0.93 |
Estimate your own workload
Input tokens/request: Output tokens/request: Requests:
| Provider | Estimated cost (USD) |
|---|
Self-host / quantized builds
Open weights under Apache 2.0 — you can run qwen3.5-9b on your own hardware instead of paying per token. Quantized to int4 it needs roughly 5.4 GB of VRAM — within reach of a single consumer GPU or a Mac with unified memory. Formats: GGUF for llama.cpp / Ollama / LM Studio, AWQ and GPTQ for GPU serving, MLX for Apple silicon.
- GGUF builds llama.cpp · Ollama · LM Studio
- AWQ builds GPU serving (vLLM)
- GPTQ builds GPU serving
- MLX builds Apple silicon
- Ollama library one-command local run
Links search the live model hubs (Hugging Face, Ollama) so they track new builds as they appear — we don't host weights. Which formats exist depends on what the model owner and community have published.
→ See VRAM, GPU and cost vs API for qwen3.5-9b, or the full self-host break-even calculator.