Ling-1T ling-1t
inclusionai · ling flagship
#78 overall
- params
- 1000B (A50B)
- arch
- moe
- license
- mit open weights
- released
- Oct 2025
- train compute
- 6.0e+24 FLOP
# architecture
- layers
- 80
- d_model
- 8192
- heads
- 64q / 8kv
- head dim
- 128
- experts
- top-8 of 256 + 1 shared
- vocab
- 157184
- family
- bailing_moe
signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.
# why ranked
Overall: #78 — score 29.1 ● low — a single signal; treat with caution
| signal | weight | input (0–100) |
|---|---|---|
| aa_intelligence | 0.70 | 25.6 |
| bench_composite | 0.30 | 37.2 |
benchmark panel evidence:
| complete panel | score (0–100) |
|---|---|
| aa-math-v1 | 37.2 |
# trend
methodology changed during this history window; score movement across that boundary is not model movement. See methodology v6.
# benchmarks
| benchmark | score | source | date |
|---|---|---|---|
| AA Intelligence Index | 9.2 | Artificial Analysis | — |
| AA Math Index | 71.3 | Artificial Analysis | — |
# where to run
no known API providers — download the weights from Hugging Face and run them yourself.
source aliases
- epoch
Ling-1T