Compare / Kolibri-1 vs Qwen3-Next 80B-A3B Thinking
Kolibri-1 vs Qwen3-Next 80B-A3B Thinking
Updated November 2026
Verdict
Kolibri-1 scores higher on the Small LM Index: 68.3 vs 63.3, although their likely ranges overlap, so the lead is not certain. On the 5 benchmarks both labs report, Kolibri-1 wins 4 and Qwen3-Next 80B-A3B Thinking wins 1. Both need about 50.8 GB of memory at 4-bit quantization. Neither model has a hosted API on OpenRouter, so you would run both yourself.
Pick Kolibri-1 if…
- you want the strongest model overall
- you mostly write code
- you need math and reasoning
- you need precise instruction following
- you work with long documents
Pick Qwen3-Next 80B-A3B Thinking if…
No clear advantage in the data we have.
Specifications
| Kolibri-1 | Qwen3-Next 80B-A3B Thinking | |
|---|---|---|
| Small LM Index | 68.3 (likely 67–70) | 63.3 (likely 59–67) |
| Lab | Aleph Alpha · Europe | Alibaba (Qwen) · China |
| Released | 2026-10-03 | 2025-09-11 |
| Parameters | 78.1B (3.5B active, MoE) | 80B (3B active, MoE) |
| Memory at 4-bit | ~49.6 GB | ~50.8 GB |
| Context window | 1M tokens | 262K tokens |
| Input | text | text |
| Reasoning | Always thinks | Always thinks |
| License | Apache-2.0 | Apache-2.0 |
| API price / 1M tokens | Self-host only | Self-host only |
| Hugging Face downloads (30 days) | 22.8K | 43.8K |
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 | Kolibri-1 | Qwen3-Next 80B-A3B Thinking |
|---|---|---|
| MMLU-Pro Knowledge | 80.0 | 82.7 |
| GPQA Diamond Knowledge | 84.3 | 77.2 |
| Humanity's Last Exam Reasoning | 21.5 | ≈ 13 |
| AIME Math | 96.0 | 87.8 |
| LiveCodeBench Coding | 85.9 | 68.7 |
| SWE-bench Verified Coding | 66.4 | ≈ 51 |
| Terminal-Bench Agentic | 27.7 | — |
| τ²-Bench Agentic | 80.4 | — |
| IFBench Instruction following | 78.1 | 61.5 |
| BFCL Agentic | 61.4 | — |
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, Kolibri-1 or Qwen3-Next 80B-A3B Thinking?
Kolibri-1 scores higher on the Small LM Index: 68.3 vs 63.3, although their likely ranges overlap, so the lead is not certain. On the 5 benchmarks both labs report, Kolibri-1 wins 4 and Qwen3-Next 80B-A3B Thinking wins 1.
Which is cheaper to run?
Neither model has a hosted API on OpenRouter, so you would run both yourself. Both need about 50.8 GB of memory at 4-bit quantization.
Which has the longer context window?
Kolibri-1 has the longer context window: 1M tokens against 262K.
Can I use Kolibri-1 and Qwen3-Next 80B-A3B Thinking commercially?
Both are released under the Apache-2.0 license. Apache 2.0 and MIT allow commercial use; other licenses can add conditions, so read the license before shipping a product.