Skip to content
aiqbench

Alibaba · model passport

Qwen: Qwen3.5-9B

qwen3-5-9b
  • Open weights
  • 9.7B params
  • 262,144 context
  • Released 2026-02-27(HF repo created)
4/27axes covered
Best percentile
76.2IFBench (Artificial Analysis)
Cohort standing
47.9median 57.4

What the data says

Benchmarked on 4 of 27 tracked axes. Every figure below is the model's own published result on that board.

Weakest axes

  • 4.5 pct — Terminal-Bench (9.2%)

Uneven: 71.7 percentile points between its best and worst axis — 76.2 pct on IFBench (Artificial Analysis) versus 4.5 pct on Terminal-Bench.

  • IFBench (Artificial Analysis)76.2 pct66.7%Verify →
  • GPQA Diamond60.2 pct79.0%Verify →
  • AIME (math)44 pct61.7%Verify →
  • Terminal-Bench4.5 pct9.2%Verify →
  • OverallNot benchmarked
  • WebDev ArenaNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • CodingNot benchmarked
  • Instruction followingNot benchmarked
  • Hard promptsNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

RepositoryQwen/Qwen3.5-9B
Parameters9.7B
ArchitectureQwen3_5ForConditionalGeneration

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Qwen/Qwen3.5-9B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Use it hosted

API model IDqwen/qwen3.5-9b
Context length262,144 tokens
Max output32,768 tokens
Acceptstext, image, video
Input price$0.10 per 1M tokens
Output price$0.15 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasLogprobsMax tokensMin-pPresence penaltyReasoningRepetition penaltyResponse formatSeedStop sequencesStructured outputsTemperatureTool choiceTool callingTop-kTop logprobsTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen/qwen3.5-9b","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • qwen3.5-9B stands at 47.9 (raw mean 46.2) — measured on 4 of 27 axes.

    standingtemplateVerify →

  • qwen3.5-9B stands at 47.9 versus a cohort median of 57.4 across 108 models released within 180 days — below the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →