Qwen2.5 Coder 32B qwen2.5-coder-32b

qwen · qwen coding

unranked — no quality signals yet

params
32.8B
arch
dense
context
32k
license
apache-2.0 open weights
released
Nov 2024
downloads/30d
552.4k

# architecture

attn shared GQA 40q/8kv feed-forward (dense) every parameter, every token out ×64 layers ctx 32,768 32.8B — no routing
layers
64
d_model
5120
heads
40q / 8kv
head dim
128
vocab
152064
family
qwen2

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: provisional — score 24.8 ● 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.

signalweightinput (0–100)
aa_intelligence 0.70 —
bench_composite 0.30 24.8

benchmark panel evidence:

complete panelscore (0–100)
aider-polyglot-v121.8
swe-bench-verified-v127.9

missing signals are dropped and the remaining weights renormalized — never imputed.

Coding: #7 — score 19.8 ● medium — one signal missing, weights renormalized

signalweightinput (0–100)
aa_coding 0.30 —
aider_polyglot 0.30 18.2
swe_bench_verified 0.40 21.1

missing signals are dropped and the remaining weights renormalized — never imputed.

Agentic: #18 — score 21.1 ● low — a single signal; treat with caution

signalweightinput (0–100)
swe_bench_verified 0.60 21.1
swe_rebench 0.40 —

missing signals are dropped and the remaining weights renormalized — never imputed.

full methodology

# trend

OGM score · last 106 days
OGM score over 106 days

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

# benchmarks

benchmarkscoresourcedate
Aider Polyglot 16.4 Aider Dec 2024
SWE-bench Verified 47.0 SWE-bench Jun 2025

# where to run

providerquantctx$/M in$/M out$/M cacheprice srctpsuptime✓
Featherless — ——— via hfrouter — —
DeepInfra 16k $0.07$0.16$0.07 via requesty — —
Nscale 128k $0.06$0.20— via hfrouter 47 —
Hyperbolic 32k $0.12$0.30— via litellm — —
Cloudflareunknown 32k $0.66$1.00— via openrouter — 100.0%
Together 16k $0.80$0.80— via litellm — —

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
aider
Qwen2.5-Coder-32B-Instruct
openrouter
qwen/qwen-2.5-coder-32b-instruct
swebench
Qwen2.5-Coder-32B-Instruct