GLM-4.7 Flash glm-4.7-flash

zai-org · glm efficient

#51 overall

params
31.2B
arch
dense
context
200k
license
mit open weights
released
Jan 2026
reasoning
yes
downloads/30d
953.7k

# architecture

attn shared · reasoning MHA ×20 feed-forward (dense) every parameter, every token out ×47 layers ctx 200,000 31.2B — no routing
layers
47
d_model
2048
heads
20q / 20kv
head dim
102
experts
top-4 of 64 + 1 shared
vocab
154880
family
glm4_moe_lite

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: #51 — score 53.7 ● high — all signals present

signalweightinput (0–100)
aa_intelligence 0.70 53.7
bench_composite 0.30 53.9

benchmark panel evidence:

complete panelscore (0–100)
aime-2025-v171.5
browsecomp-v136.3

full methodology

# trend

OGM score · last 106 days
OGM score over 106 days
overall rank (up is better)
Overall rank over 106 days

methodology changed during this history window; score movement across that boundary is not model movement. See methodology v6.

# benchmarks

benchmarkscoresourcedate
AA Intelligence Index 14.9 Artificial Analysis —
AIME 2025 0.9 / 1 LLM Stats —
Browsecomp 0.4 / 1 LLM Stats —

# where to run

providerquantctx$/M in$/M out$/M cacheprice srctpsuptime✓
Featherless — ——— via hfrouter — —
Z.AI — ——— via hfrouter 61 —
DeepInfra 198k $0.06$0.40— via hfrouter 20 —
Cloudflareunknown 128k $0.06$0.40— via openrouter — 100.0%
Amazon Bedrock — $0.07$0.40— bedrock — — ✓
Novitabf16 200k $0.07$0.40— novita — 2.2% ✓

sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price

source aliases
aa
glm-4.7-flash
openrouter
z-ai/glm-4.7-flash