whatmodel.sh / GLM-5

GLM-5

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

GLM-5 is an open-weights language model from Z.ai, released February 2026. Its published size is 744B parameters (40B active per token) and a 205k-token context window. Licence: mit.

69.8±0.6#86 of 719 LLMsfrom 59 benchmark results across 17 sources

$ how it compares

The LLMs ranked either side of GLM-5, out of 828 with a score, then more from Z.ai.

#modelscorewhereRAMreleased
91MiMo-V2-Pro Xiaomi70±1comparison only2026-03
92Kimi K2.5 Moonshot AI70±1api $1.20/M +local~160GB2026-01
93GLM-5 Z.ai70±1api $1.55/M +local~119GB2026-02
94Qwen3-Max (2026-01-23) Alibaba70±1comparison only2026-01
95MiMo-V2.5-Pro (1T-A42B) Xiaomi70±1api $0.65/M +local~163GB2026-04
more from Z.ai
33GLM-5.3 Z.ai76±1api $2.15/M +local~239GB2026-08
52GLM-5.3-Flash Z.ai74±1api $0.24/M +local~103GB2026-08
57GLM-5.2 Z.ai74±1api $2.15/M +local~121GB2026-06

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #72 of 300 · 38% of its score · 18 tests

Beats GPT-5 Pro, GPT-5, Gemini 3 Flash (Preview) here, though they rank higher overall. Trails MiMo-V2.5 (310B-A15B), Muse Glimmer 30B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
SWE-bench Verified +172.1% ✎▁··▁·▁▂▁▄█▄▁ beats 10/29hard · avg 19%▇▇▇ 8.3%sharp · at its level · several runs
MCP-Atlas (public set) lab +267.8% ✎▄······▄▄▄▄█ beats 5/6hard · avg 25%▇▇ 5.2%at its level · several runs
Terminal-Bench 2.0 lab +156.2% ✎▂▂▅▅████▆▂▃▃ beats 23/36very hard · avg 14%▇ 3.0%at its level · small run · several runs
BrowseComp lab +162.0% ✎▂·▄▁▂▃▆▄█▂▃▄ beats 27/42hard · avg 22%▇ 2.2%at its level · small run · several runs
SWE-bench Pro (split unspecified) lab55.1% ✎▃·▁▁▁▂▂▃█▇▃▁ beats 31/55very hard · avg 9.6%▇ 2.0%sharp · at its level · small run
τ³-bench banking_knowledge (v1.0.1 grading)9.8% ✎▃▃▂▃·▆▃▂█▅▅▂ beats 1/25very hard · avg 4.1%▇ 1.8%typical
Toolathlon (Tool Decathlon) lab +138.0% ✎▂▂▄▂▄▄▆▂█▆██ beats 10/27very hard · avg 12%▇ 1.6%at its level · small run · several runs
BrowseComp with context management lab +175.9% ✎▅·▅▃·█▃·█▅·▃ beats 5/14medium · avg 43%▇ 1.5%small run · several runs
10 more tests
testresultvs the others in its rundifficultycountswhy
τ³-bench airline (tau2-bench ≥ v1.0.1)82.5% ✎▃······▃·▃█▃ beats 3/6medium · avg 49%▇ 1.5%typical
τ²-bench aggregate lab89.7% ✎▁▁▁·▂▂▄▃▃▅█▃ beats 36/37medium · avg 48%▇ 1.5%small run
Vibe Code Bench (test split, 50 web apps)23.4% ✎▃··▃·▃██▃▃▃█ beats 8/14very hard · avg 2.3%▇ 1.4%sharp · at its level · noisy
BrowseComp-ZH lab72.7% ✎▁▁▁▁▁▃·▁▄▅▄█ beats 23/23hard · avg 22%▇ 1.4%at its level · small run
SWE Atlas (Codebase Q&A, Test Writing, Refactoring; mini-SWE-agent) pass@124.0% ✎▄·▄··█···▄·█ beats 3/6very hard · avg 6.1%▇ 1.4%typical
τ³-bench retail (tau2-bench ≥ v1.0.1)73.7% ✎▄▄·█··▄·▄··▄ beats 1/6medium · avg 46%▇ 1.3%typical
τ³-bench telecom (tau2-bench ≥ v1.0.1)86.8% ✎█·▄█▄······▄ beats 1/6medium · avg 62%▇ 1.2%typical
SWE-bench Multilingual lab73.3% ✎▁··▁▃▂▅▇█▃▃▁ beats 39/60hard · avg 28%▇ 1.2%at its level · small run
Humanity's Last Exam with tools lab50.4% ✎▂▂▂▁▄▂▃▆█▇▂▆ beats 31/54hard · avg 16%▇ 0.8%at its level · small run
τ³-bench lab69.2% ✎▁▁▂▁·▁▁·▁▁▂█ beats 18/25medium · avg 36%▇ 0.6%blunt · at its level · small run
coding▁▁▁▃▆██▇▆▄▃▁ #70 of 458 · 20% of its score · 6 tests

Beats DeepSeek V4 Flash Vision (Exp), GPT-5.2, GPT-5 here, though they rank higher overall. Trails Gemini 3.5 Flash-Lite, MiMo-V2.5 (310B-A15B), MiMo-V2.5-Pro (1T-A42B) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
SWE-bench Verified +172.1% ✎▁··▁·▁▂▁▄█▄▁ beats 10/29hard · avg 19%▇▇▇ 8.3%sharp · at its level · several runs
Arena WebDev (Code Arena)1434.0 ✎▁▂▄▃▅█▄▇▆▄▁▁ beats 53/119avg 981▇▇ 5.3%sharp
SWE-bench Pro (split unspecified) lab55.1% ✎▃·▁▁▁▂▂▃█▇▃▁ beats 31/55very hard · avg 9.6%▇ 2.0%sharp · at its level · small run
Vibe Code Bench (test split, 50 web apps)23.4% ✎▃··▃·▃██▃▃▃█ beats 8/14very hard · avg 2.3%▇ 1.4%sharp · at its level · noisy
SWE Atlas (Codebase Q&A, Test Writing, Refactoring; mini-SWE-agent) pass@124.0% ✎▄·▄··█···▄·█ beats 3/6very hard · avg 6.1%▇ 1.4%typical
SWE-bench Multilingual lab73.3% ✎▁··▁▃▂▅▇█▃▃▁ beats 39/60hard · avg 28%▇ 1.2%at its level · small run
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #44 of 393 * · 3.0% of its score · 1 test

Beats MiMo-V2.6-Flash (309B-A15B), GPT-6 Luna, GPT-5.6 Luna here, though they rank higher overall. Trails MiMo-V2.5-Pro (1T-A42B) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
LMArena (text)1457.6 ✎▁▁▁▂▃▃▃▆▆▇█▃ beats 329/379avg 1268▇ 3.0%blunt
math▁▂▃▅▆▅▄█▄▁▁▁ #70 of 398 · 23% of its score · 13 tests

Beats Claude Sonnet 4.6, Claude Opus 4.5, Qwen3.7 Plus here, though they rank higher overall. Trails INTELLECT-3, EXAONE 4.5 33B, Qwen3 235B-A22B Thinking 2507 here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
MathArena ArXivMath February 2026 (32 problems)41.4% ✎▁▁▁▁█▇▄▃▁▃·▃ beats 14/24very hard · avg 9.1%▇ 3.7%at its level
MathArena ArXivMath March 2026 (30 problems)45.8% ✎█▃▅·█▃▃·▅·▃▃ beats 2/14very hard · avg 15%▇ 3.5%at its level
MathArena ArXivMath January 2026 (23 problems)54.3% ✎▂··▄▂▄█▆▆▄▄█ beats 9/23hard · avg 22%▇ 2.8%at its level
AIME 2026 lab95.4% ✎▁▁▁▂▁▂▆█▆▂·▁ beats 41/49medium · avg 61%▇ 2.1%sharp · maxed out · small run
HMMT February 2026 lab82.8% ✎▃▃▂▃·██▆█▅▃▅ beats 22/33hard · avg 30%▇ 2.1%sharp · small run
MathArena ArXivMath December 2025 (17 problems)38.2% ✎▃·▃▃█▁·▄·▁▁▁ beats 5/18hard · avg 19%▇ 1.8%at its level
IMO-AnswerBench lab82.5% ✎▁▁·▁▁▁▄▅█▄▂▃ beats 28/42medium · avg 37%▇ 1.6%small run
OTIS mock AIME 2024–202580.0% ✎▃▃▄▃▃▃▇▇█▃▂▅ beats 113/193very hard · avg 8.5%▇ 1.5%sharp · noisy
5 more tests
testresultvs the others in its rundifficultycountswhy
HMMT February 2025 lab97.9% ✎▁▁▁▃▁▅▅▃██▁▄ beats 29/31easy · avg 77%▇ 1.5%maxed out · small run
MathArena BrokenArXiv March 2026 (56 problems)12.1% ✎█·▆█▂·▂▂···▂ beats 6/14very hard · avg 0.8%▇ 1.0%sharp · noisy
MathArena BrokenArXiv February 2026 (31 problems)11.7% ✎▃▂▂▃█▃▂·▃··▂ beats 8/16very hard · avg 0.7%▇ 0.9%sharp · noisy
HMMT November 2025 lab96.9% ✎▁▁····▁▅▃█▇▃ beats 23/23easy · avg 74%▇ 0.9%maxed out · small run
AIME 2026 I lab92.7% ✎█··········█ beats 0/1—not counted yet: no other model in the same run to compare it with
web search▂▁▂▆▄█▃▃▁▁·▁ #53 of 80 · 6.0% of its score · 4 tests

Beats Kimi K2.6, GPT-5 Pro, DeepSeek V4 Flash (Preview, Apr 2026) here, though they rank higher overall. Trails GPT-4o mini (2024-07-18), Mistral Small 3.1 24B (2503), Llama 4 Scout (109B-A17B) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
BrowseComp lab +162.0% ✎▂·▄▁▂▃▆▄█▂▃▄ beats 27/42hard · avg 22%▇ 2.2%at its level · small run · several runs
BrowseComp with context management lab +175.9% ✎▅·▅▃·█▃·█▅·▃ beats 5/14medium · avg 43%▇ 1.5%small run · several runs
BrowseComp-ZH lab72.7% ✎▁▁▁▁▁▃·▁▄▅▄█ beats 23/23hard · avg 22%▇ 1.4%at its level · small run
Humanity's Last Exam with tools lab50.4% ✎▂▂▂▁▄▂▃▆█▇▂▆ beats 31/54hard · avg 16%▇ 0.8%at its level · small run
other tests16 not in a task · 30% of its score
testresultvs the others in its rundifficultycountswhy
GPQA Diamond +287.8% ✎▁▁▁▁▂▄▆▄▅▇█▅ beats 159/208hard · avg 40%▇▇▇ 8.7%sharp · several runs
Humanity's Last Exam lab +130.5% ✎▁▁▁▃▆▆▇██▆▁▃ beats 72/104very hard · avg 5.5%▇ 2.7%at its level · small run · several runs
ARC-AGI-2 public evaluation set (ARC Prize model_baseline, pass@2)5.4% ✎█··▂▅▁▃▁▃·▁▃ beats 12/24very hard · avg 0.3%▇ 2.4%sharp · noisy
ARC-AGI-1 public evaluation set (ARC Prize model_baseline, pass@2)58.6% ✎▁·▇▃▁·█▇▁·▇▃ beats 11/27very hard · avg 4.1%▇ 2.3%sharp · at its level · noisy
KINA (Knowledge Index of Noah's Ark) accuracy, 899 ten-option items35.9% ✎▁▁▂▁▃▄█▆▄▅▂▃ beats 28/39very hard · avg 22%▇ 2.2%blunt · at its level
EQ-Bench 3 (emotional intelligence, judge Elo)1307.3 ✎▁·▂▁▄▅▅▃█▄▁▁ beats 49/62avg 1044▇ 1.8%typical
WildClawBench42.6% ✎▃·▅▅█·▅█▅▃▃▃ beats 10/17very hard · avg 11%▇ 1.6%at its level
Chess Puzzles (Epoch)10.0% ✎█▁▁▁▃▂▁▄▂▂▁▁ beats 55/135very hard · avg 5.9%▇ 1.6%sharp · noisy
8 more tests
testresultvs the others in its rundifficultycountswhy
LongBench v2 lab64.5% ✎▂·▁·▂▂▁▃▄▁█▄ beats 27/28very hard · avg 35%▇ 1.4%at its level · small run
CyberGym lab +148.3% ✎▄···▄··▄█▄█▄ beats 1/8hard · avg 19%▇ 1.3%at its level · small run · several runs
AraGen-v3 3C3H score (Inception, Claude Sonnet 4 judge)46.0% ✎▁▁·▁▂▂▄█▄▅▄▄ beats 40/66hard · avg 31%▇ 1.2%typical
LongSocialBench (full-context, 4-option MCQ over discussion threads)41.5% ✎▆▂·▄█·▂▂·▄▄▂ beats 7/16very hard · avg 36%▇ 1.1%blunt · at its level
NL2Repo lab35.9% ✎▂▂·▂▂▄█▆▆▂▄▆ beats 8/21very hard · avg 10%▇ 1.1%at its level · small run
SOOHAK Challenge (340 problems, Avg@3)9.6% ✎█·█████··█·█ beats 2/7very hard · avg 4.1%▇ 0.5%typical
SOOHAK-Mini (702 problems, Avg@3)63.1% ✎▄··▄·▄·▄▄█·▄ beats 4/7medium · avg 36%▇ 0.4%blunt · at its level
Vectara hallucination leaderboard (HHEM-2.3, factual consistency rate)89.9% ✎▁▂▁▂▅█▂▃▄▃▁▁ beats 44/98very easy · avg 90%▇ 0.0%blunt

$ where to run it

Through an API

api · $1.55/M · $1 in / $3.2 out per 1M tokens · 205k context · docs.z.ai ↗ · price sourceStandard base token rate from models.dev. Provider-specific tiers, long-context surcharges, caching and promotions may differ.

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run GLM-5 · needs ~446.9 GB · from huggingface.co/unsloth
    chat in the terminal
    llama cli -hf unsloth/GLM-5-GGUF:UD-Q4_K_XL
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf unsloth/GLM-5-GGUF:UD-Q4_K_XL

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run GLM-5 · needs ~446.9 GB · from huggingface.co/unsloth

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/unsloth/GLM-5-GGUF@UD-Q4_K_XL -y
  4. Then run lms ls to see its name, and lms chat followed by that name

MLX (Mac)

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run GLM-5 · needs ~418.6 GB, Mac only · from huggingface.co/mlx-community
    chat in the terminal
    mlx_lm.chat --model mlx-community/GLM-5-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model mlx-community/GLM-5-4bit

Weights and documentation: huggingface.co/zai-org ↗

$ how the score was worked out · zai-org/GLM-5.1 model card 18%, MathArena ArXivMath / BrokenA… 14%, GLM-5 Team (Zhipu AI) — GLM-5… 14% and 14 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
zai-org/GLM-5.1 model card · lab-reported13▇▇▇▇▇ 18%mit
MathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasets6▇▇▇▇ 14%CC BY-SA 4.0
GLM-5 Team (Zhipu AI) — GLM-5 technical report · lab-reported10▇▇▇▇ 14%CC BY 4.0
Epoch AI — internal evaluations4▇▇▇▇ 13%CC BY 4.0
zai-org/GLM-5 model card · lab-reported8▇▇▇ 11%MIT
LMArena leaderboard dataset2▇▇▇ 8%CC BY 4.0
τ²/τ³-bench leaderboard: Sierra's own evaluation runs4▇▇ 6%MIT
ARC Prize Foundation model_baseline outputs on ARC-AGI-1/2 public evaluation sets2▇ 5%MIT
Knowledge Index of Noah's Ark (arXiv:2606.05104v2)1▇ 2%CC BY 4.0
EQ-Bench 3 canonical leaderboard results1▇ 2%MIT
WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation (arXiv:2605.10912v1)1▇ 2%CC BY 4.0
Vibe Code Bench (Vals AI, arXiv:2603.04601v3)1▇ 1%CC-BY-4.0
SWE Atlas: Benchmarking Coding Agents Beyond Issue Resolution (arXiv:2605.08366v1)1▇ 1%CC BY 4.0
Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)1▇ 1%Apache-2.0
LongSocialBench: Do Long-Context LLMs Understand Online Discussion Threads? (arXiv:2610.04118v1)1▇ 1%CC BY 4.0
Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs — authors' own model runs2▇ 1%CC BY-SA 4.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)1▇ 0%Apache-2.0
prior · pull toward the middle▇ 0%
weighted average= 69.8

not counted (1)

zai-org/GLM-5 model cardAIME 2026 I [correct]92.7%fewer than three models have this test yet, or this source ran it for this model alone

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json