inclusionAI: Ling 3.1 Flash
inclusionAI
Capabilities
Strengths
- Free inference with no token cost
- Extended context window (262K tokens) for long documents
- Hybrid reasoning with extended thinking capability
- MoE efficiency reduces latency compared to dense flagships
Weaknesses
- Very slow inference speed (0.14 speed_score)
- Limited provider transparency and disclosure
- Reasoning capability moderate (0.7), not specialized
- No vision or multimodal support
Pricing
Input / 1M tokens
$0
Output / 1M tokens
$0
Context window
262.144k
Transparency
Open weights
10.0 / 10
Open training data
0.0 / 10
Open methodology
6.0 / 10
Licence openness
7.0 / 10
Provider disclosure
5.0 / 10
FMTI company score
N/A
Open-weight MoE model with accessible methodology for hybrid reasoning architecture. Limited public disclosure on training data composition or provider infrastructure details. Licence openness reflects standard open-weight licensing.
Sustainability
Inference energy
N/A
Training footprint
N/A
Provider infrastructure
5.0 / 10
No public commitments or data available on energy efficiency or green infrastructure. Sparse MoE architecture (25B active / 560B total) may reduce inference costs relative to dense models, but provider sustainability profile unknown.