Qwen3.5 9B qwen3.5-9b

qwen · qwen efficient

#34 overall

params
9.7B
arch
dense
context
256k
license
apache-2.0 open weights
released
Feb 2026
reasoning
yes
downloads/30d
8.4M

# why ranked

Overall: #34 — score 18.6 ● medium — one signal missing, weights renormalized

signalweightinput (0–100)
aa_intelligence 0.50 21.4
arena_elo 0.30
bench_composite 0.20 11.5

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

Coding: provisional — score 27.6 ● 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_coding 0.25 27.6
aider_polyglot 0.25
arena_elo_code 0.15
swe_bench_verified 0.35

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

full methodology

# trend

collecting history — trends appear after a few daily runs.

# benchmarks

benchmarkscoresourcedate
AA Coding Index 28.7 Artificial Analysis
AA Intelligence Index 21.4 Artificial Analysis
Aa-lcr 0.6 / 1 LLM Stats
Bfcl-v4 0.7 / 1 LLM Stats
C-eval 0.9 / 1 LLM Stats
Deep-planning 0.2 / 1 LLM Stats

# where to run

providerquantctx$/M in$/M out$/M blendtpsuptime
DeepInfrabf16 256k $0.10$0.15$0.11 99.5%
SiliconFlowfp8 256k $0.10$0.15$0.11 99.6%
Venicefp8 250k $0.10$0.15$0.11 99.6%
Togetherunknown 256k $0.17$0.25$0.19 99.9%

blended = (3·input + output) / 4 per 1M tokens · cheapest first

source aliases
aa
qwen3.5-9b
openrouter
qwen/qwen3.5-9b