Compare / Ornith 1.5 9B vs Step3-VL-10B
Ornith 1.5 9B vs Step3-VL-10B
Updated November 2026
Verdict
Ornith 1.5 9B scores higher on the Small LM Index: 72.9 vs 53.3. On the 2 benchmarks both labs report, Ornith 1.5 9B wins 2 and Step3-VL-10B wins 0. Both need about 8 GB of memory at 4-bit quantization. Neither model has a hosted API on OpenRouter, so you would run both yourself.
Pick Ornith 1.5 9B if…
- you want the strongest model overall
- you work with long documents
- you want the more widely used model (more quantized builds and community support)
Pick Step3-VL-10B if…
No clear advantage in the data we have.
Specifications
| Ornith 1.5 9B | Step3-VL-10B | |
|---|---|---|
| Small LM Index | 72.9 (likely 67–78) | 53.3 (likely 40–64) |
| Lab | Ornith · Other | StepFun · China |
| Released | 2026-08-18 | 2026-01 |
| Parameters | 9.7B (dense) | 10B (dense) |
| Memory at 4-bit | ~6.8 GB | ~8 GB |
| Context window | 262K tokens | 64K tokens |
| Input | text, image | text, image |
| Reasoning | Always thinks | Always thinks |
| License | MIT | Apache-2.0 |
| API price / 1M tokens | Self-host only | Self-host only |
| Hugging Face downloads (30 days) | 582.5K | — |
Benchmarks
Higher is better. Bold marks the winner where both labs report a score. ≈ marks the index model's estimate for a score that wasn't reported, ✓ an independent measurement and * a score measured by another lab.
| Benchmark | Ornith 1.5 9B | Step3-VL-10B |
|---|---|---|
| GPQA Diamond Knowledge | 86.4 | 69.0✓ |
| Humanity's Last Exam Reasoning | 20.2 | 10.8✓ |
| SWE-bench Verified Coding | 70.6 | ≈ 48 |
| Terminal-Bench Agentic | 46.2 | — |
| τ²-Bench Agentic | — | 16.1✓ |
| IFBench Instruction following | — | 50.2✓ |
| AA-LCR Long context | — | 0.0✓ |
Scores are mostly reported by each lab with its own settings, so small gaps are noise. See the methodology. Open in the interactive comparison to add more models.
Questions
Which is better, Ornith 1.5 9B or Step3-VL-10B?
Ornith 1.5 9B scores higher on the Small LM Index: 72.9 vs 53.3. On the 2 benchmarks both labs report, Ornith 1.5 9B wins 2 and Step3-VL-10B wins 0.
Which is cheaper to run?
Neither model has a hosted API on OpenRouter, so you would run both yourself. Both need about 8 GB of memory at 4-bit quantization.
Which has the longer context window?
Ornith 1.5 9B has the longer context window: 262K tokens against 64K.
Can I use Ornith 1.5 9B and Step3-VL-10B commercially?
Ornith 1.5 9B is released under MIT and Step3-VL-10B under Apache-2.0. Apache 2.0 and MIT allow commercial use; other licenses can add conditions, so read the license before shipping a product.