Rank #10 / 13
Qwen3.8-27B
Qwen3.8-27B is Alibaba's Qwen model (open weights). It is ranked 10 of 13 models with enough coverage on the Visual Intelligence Index.
- Alibaba
- Open weights
- Qwen
- Video input
- Audio
- Structured output
- Tool calling
- Streaming
- Self-hostable
Comparison Summary
Qwen3.8-27B ranks 10 of 13 on the perception index. Immediately ahead is Gemini 3.1 Pro. Immediately behind is Qwen3.5-35B-A3B. List price is $0.25 per million input tokens and $1.00 per million output. 3 of 13 listed models cost less to prompt. Latency is 15.7s with a 256K context window and — tokens/s. Prices and most scores in this snapshot are placeholders, not sourced figures. Strongest capability in this set is Causal Reasoning (71%). Weakest is Action / Event Understanding (64%).
- Visual Intelligence
- 67.5%
- Availability
- Open weights
- Input / 1M tokens
- $0.25
- Output / 1M tokens
- $1.00
- Latency
- 15.7s
- Context Window
- 256K
- Speed (tokens/s)
- —
- Parameters
- 27B
- Time to first token
- —
All benchmark scores
| Benchmark | Metric | Score | 95% CI | Setting | Source |
|---|---|---|---|---|---|
| Video-MME v2 | Accuracy, no subtitles | 50.5% | 48.8–52.3 | 1 fps, 768-frame cap (226 videos thinned) | Reported |
| LVBench | Test accuracy | 64.6% | 62.2–67.0 | 1 fps, 768-frame cap (103 videos thinned) | Reported |
| Perception Test | Overall accuracy | 72.3% | 68.1–76.4 | 1 fps | Reported |
| NExT-QA | Accuracy (hard split) | 82.2% | 81.1–83.2 | 1 fps | Reported |
| Q-Bench Video | Overall accuracy | 66.7% | 63.5–69.9 | 1 fps | Reported |
| EgoSchema | Accuracy (fullset) | 79.1% | 75.6–82.7 | 1 fps | Reported |
| UCF101-AD | Accuracy | 45.5% | 44.0–47.0 | 1 fps | Reported |
Benchmark Scores
Compare reported model scores across each available benchmark or capability index.
13 of 13 models
Display
Capabilities Index Scores
Compare reported model scores across each available benchmark or capability index.
13 of 13 models
Display
Usability
4.5/ 5
What it's like to build against this model, scored out of five from the developer-facing capabilities in the dataset.
- Reachable1.0
There is an endpoint you can call without hosting anything.
- First-party API — supported
- Third-party API — supported
- Portable1.0
You can run it yourself, and are not tied to one vendor.
- Open weights — supported
- Self-hostable — supported
- Multimodal input1.0
It takes the footage directly, rather than frames you extracted.
- Video ingestion — supported
- Audio — supported
- Programmable1.0
Output you can parse, and tools it can call on its own.
- Structured output — supported
- Tool calling — supported
- Operable0.5
Usable interactively and in bulk, not only one call at a time.
- Streaming — supported
- Batch API — not supported
Other Alibaba models
Other models from the Alibaba family.
| Date | Model | Visual Intelligence | Latency |
|---|---|---|---|
| Jan 28, 2026 | Qwen3.5-35B-A3B | 64.7% | 9.4s |
| Jul 22, 2026 | Qwen3.8-27B | 67.5% | 15.7s |
| Aug 3, 2026 | Qwen3.8-Max | 85.6% | 10.8s |