K2 Horizon 375B A23B k2-horizon-375b-a23b
IFM · ifm
#11 overall
- params
- 379.2B
- arch
- moe
- license
- apache-2.0 open weights
- released
- Sep 2026
- downloads/30d
- 16.2k
# architecture
- layers
- 61
- d_model
- 6144
- heads
- 48q / 8kv
- head dim
- 128
- experts
- top-8 of 192 + 1 shared
- vocab
- 250624
- family
- k2_horizon
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: #11 — score 88.5 ● low — a single signal; treat with caution
| signal | weight | input (0–100) |
|---|---|---|
| aa_intelligence | 0.70 | 86.2 |
| bench_composite | 0.30 | 94.0 |
benchmark panel evidence:
| complete panel | score (0–100) |
|---|---|
| aa-coding-v1 | 94.0 |
Coding: provisional — score 83.8 ● low — a single signal; treat with caution
provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.
| signal | weight | input (0–100) |
|---|---|---|
| aa_coding | 0.30 | 83.8 |
| aider_polyglot | 0.30 | — |
| swe_bench_verified | 0.40 | — |
missing signals are dropped and the remaining weights renormalized — never imputed.
# 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 Coding Index | 61.5 | Artificial Analysis | — |
| AA Intelligence Index | 30.5 | Artificial Analysis | — |
# where to run
no known API providers — download the weights from Hugging Face and run them yourself.
source aliases
- aa
k2-horizon-375b-a23b