Skip to content
aiqbench

Zai · model passport

Z.ai: GLM 5

glm-5
  • Open weights
  • MIT licence
  • 753.9B params
  • 204,800 context
  • Released 2026-02-11(HF repo created)
16/27axes covered
Overall · LMArena
1,457#56 of 396
Best percentile
86.4IFBench (Artificial Analysis)
Cohort standing
62.6median 55.9

What the data says

Benchmarked on 16 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 86.4 pct — IFBench (Artificial Analysis) (72.3%)
  • 86.1 pct — Overall (1,457)
  • 85.1 pct — Hard prompts (1,478)

Uneven: 52 percentile points between its best and worst axis — 86.4 pct on IFBench (Artificial Analysis) versus 34.4 pct on SWE-bench Verified.

  • IFBench (Artificial Analysis)86.4 pct72.3%Verify →
  • Overall86.1 pct1,457Verify →
  • Hard prompts85.1 pct1,478Verify →
  • Instruction following84.6 pct1,446Verify →
  • Coding81.1 pct1,498Verify →
  • LiveBench78.7 pct68.9%Verify →
  • GPQA Diamond74.7 pct87.8%Verify →
  • Factuality Arena67.4 pct1,452Verify →
  • Terminal-Bench65.9 pct52.4%Verify →
  • SimpleBench57.5 pct53.2%Verify →
  • AIME (math)57 pct80.0%Verify →
  • FrontierMath53.6 pct16.4%Verify →
  • WebDev Arena48.2 pct1,435Verify →
  • HHEM Hallucination (Vectara)44.8 pct89.9%Verify →
  • ARC-AGI-234.8 pct4.9%Verify →
  • SWE-bench Verified34.4 pct72.1%Verify →
  • Vision ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • Humanity's Last ExamNot 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

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

Repositoryzai-org/GLM-5
Parameters753.9B
ArchitectureGlmMoeDsaForCausalLM
LicenceMIT

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "zai-org/GLM-5"
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 IDz-ai/glm-5
Context length204,800 tokens
Max output131,072 tokens
Input price$1 per 1M tokens
Cached input price$0.20 per 1M tokens
Output price$3.20 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogprobsMax tokensPresence 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":"z-ai/glm-5","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.

  • glm-5 stands at 62.6 (raw mean 65) — measured on 16 of 27 axes.

    standingtemplateVerify →

  • glm-5 stands at 62.6 versus a cohort median of 55.9 across 106 models released within 180 days — above 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 →