Qwen3.5 4B
Qwen3.5 4B is a small multimodal model in Alibaba's Qwen3.5 series, taking text, image and video input.
Model card on Hugging Face- Input / output price
- $0.04 / $0.07 per 1M
- Context
- 262K tokens
- Released
- Mar 2, 2026
Paste into Claude Code, Cursor, or Codex
Help me set up the Respan gateway: fetch and follow https:/
Providers
One provider serves this model, priced per 1M tokens after gateway discounts.
| Provider | Input | Output |
|---|---|---|
EmpirioLabs empiriolabs/qwen3-5-4b | $0.04/M | $0.07/M |
Pricing
USD per 1M tokens on EmpirioLabs, after gateway discounts.
- Input
- $0.04 per 1M
- Output
- $0.07 per 1M
- Cache read
- $0.02 per 1M
Cost by volume
Total cost from 0 to 1T tokens.
Throughput and latency
30-day averages from requests through the Respan gateway. Under each figure, every model on the gateway by its average, this one in blue.
Output speed
Higher is better41tokens/sacross providers
Response latency
Lower is better6.4sacross providers
Benchmarks
The Respan Index plus 16 published scores from 2 sources, each linked to where it was published.
17.5/ 100
Respan Index#147 of 176
- 147Qwen3.5 4B17.5
- 146Kimi K217.6
- 148MiniMax M217.4
Published scores
| Benchmark | Score | Source | Published |
|---|---|---|---|
| LiveCodeBench v6huggingface.co | 55.8% | huggingface.co | Mar 2, 2026 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| TAU2-Benchhuggingface.co | 79.9% | huggingface.co | Mar 2, 2026 |
| BFCL-V4huggingface.co | 50.3% | huggingface.co | Mar 2, 2026 |
| OSWorld-Verifiedhuggingface.co | 35.6% | huggingface.co | Mar 2, 2026 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| GPQA Diamondhuggingface.co | 76.2% | huggingface.co | Mar 2, 2026 |
| SuperGPQAhuggingface.co | 52.9% | huggingface.co | Mar 2, 2026 |
| LongBench v2huggingface.co | 50% | huggingface.co | Mar 2, 2026 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| HMMT Feb 25huggingface.co | 74% | huggingface.co | Mar 2, 2026 |
| OTIS Mock AIME 2024-2025Epoch AI | 55.8% | Epoch AI | Oct 5, 2026 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| IFEvalhuggingface.co | 89.8% | huggingface.co | Mar 2, 2026 |
| MMLU-Reduxhuggingface.co | 88.8% | huggingface.co | Mar 2, 2026 |
| MMLU-Prohuggingface.co | 79.1% | huggingface.co | Mar 2, 2026 |
| MMMLUhuggingface.co | 76.1% | huggingface.co | Mar 2, 2026 |
| IFBenchhuggingface.co | 59.2% | huggingface.co | Mar 2, 2026 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| MMMUhuggingface.co | 77.6% | huggingface.co | Mar 2, 2026 |
| MMMU-Prohuggingface.co | 66.3% | huggingface.co | Mar 2, 2026 |
How Qwen compares: DeepSeek vs Qwen, Claude vs Qwen, Gemini vs Qwen
Compare Qwen3.5 4B: Qwen3.5 4B vs Qwen3.8-Max, Qwen3.5 4B vs Qwen3.8-Flash-Next, Qwen3.5 4B vs Qwen3.7-Max, Qwen3.5 4B vs Qwen3.8 27B, Qwen3.5 4B vs Qwen3.6-Max-Preview, Qwen3.5 4B vs Qwen3.7-Plus
Apps
Use Qwen3.5 4B in the tools you already run. Each guide points the app at the Respan gateway.
Setup guides for more apps are in the integration docs.
Qwen3.5 4B FAQ
Short answers, built from the data on this page.
Qwen3.5 4B is a model from Qwen, released on Mar 2, 2026. Qwen3.5 4B is a small multimodal model in Alibaba's Qwen3.5 series, taking text, image and video input.
On the Respan gateway, Qwen3.5 4B costs $0.04 per 1M input tokens and $0.07 per 1M output tokens on EmpirioLabs, after gateway discounts.
Qwen3.5 4B is served by EmpirioLabs.
Qwen3.5 4B accepts up to 262K tokens of context.
Point any OpenAI-compatible SDK at https://gateway.respan.ai/v1 with your Respan API key and set the model to "qwen3.5-4b". To set it up with a coding agent, copy the agent setup prompt at the top of this page.
Qwen3.5 4B scores 17.5 out of 100 on the Respan Index v1.1, ranking #147 of 176 models. By category: Coding 1.3, Agents 19.8, Reasoning 20.5, Knowledge & language 10.7, and Human preference 40.2 (estimated). The Respan Index combines independent and lab-reported benchmark results into one score, using benchmark data collected by Respan.
Qwen3.5 4B has 16 published benchmark scores, including LiveCodeBench v6 55.8%, TAU2-Bench 79.9%, and GPQA Diamond 76.2%. Each score links to where it was published; the data is collected by Respan.
Certified where it counts.
Built to meet the security and privacy standards that enterprise and healthcare teams require.
ISO 27001
The internationally recognized standard for information security management.
SOC 2
Secure, compliant management of your data across all of our systems.
GDPR
Operated under GDPR, the world's strictest standard for data privacy.
HIPAA
HIPAA compliant, with a BAA available for healthcare teams.