Inference
EN
Z.AI

GLM-5.1

GLM-5.1 is a reasoning model from Z.AI in the GLM-5 family. 38 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-04
RankedEnough results for an index.
IndexAll categories, out of 100.56.0±4.1Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.102/sMedian 67/s
First tokenSeconds to the first output token.0.5 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$1.40Median $0.60
Output / 1MUS dollars per 1M output tokens.$4.40Median $2.55
ContextMaximum tokens in one request.205KMedian 262K
EloLMArena rating and rank.1463±4 #40Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
z-ai/glm-5.1
Max outputMaximum output tokens per request.
131K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.26
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
11 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
ChutesHost$0.90$2.90203K
GMI CloudHost$0.98$3.08203K
Deep InfraDeprecatedHost$1.05$3.50203K
Zhipu AIMaker$0.896$3.59200K
SiliconFlowHost$1.19$3.74205K
CrusoeDeprecatedHost$1.20$4.40200K
Novita AIHost$1.38$4.40205K
FriendliHost$1.40$4.40203K
OpenRouterRouter$1.40$4.40205K
Together AIDeprecatedHost$1.40$4.40203K
Z.AIMaker$1.40$4.40200K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

753.86B parameters
16-bit
1584.1 GB
8-bit
842.0 GB
4-bit
480.9 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.13 counted59.277 / 218
CodingCode writing and repair.8 counted57.084 / 209
ReasoningLogic problems and puzzles.3 counted50.9108 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.5 counted51.7123 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted60.3—
MathMath problems.8 counted54.965 / 168

Results

38 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
τ²-Bench Tool-Agent-User EvaluationAgentic97.7%9 / 149
Artificial Analysis IFBenchInstruction76.3%13 / 146
τ³-Bench Tool-Agent-User EvaluationAgentic70.6%6 / 32
GPQA diamond · Epoch AIKnowledge89.9% ±2.141 / 217
FrontierMath-2025-02-28-PrivateMath33.4% ±2.815 / 69
AIMEMath91.9% ±0.421 / 95
Gert Labs Composite Game BenchmarkAgentic60.1%12 / 52
Terminal-Bench 2.0 · Vals AIAgentic53.9% ±5.316 / 66
OTIS Mock AIME 2024-2025Math93.3% ±3.851 / 200
MMLU ProKnowledge86.9% ±0.337 / 135
Artificial Analysis GPQA DiamondKnowledge86.8%61 / 199
FrontierMath-Tier-4-2025-07-01-PrivateMath12.5% ±4.817 / 55
Harvard-MIT Mathematics Tournament November 2025Math94.0%4 / 12
Terminal-Bench 2.0 · BenchLMAgentic63.5%14 / 41
Artificial Analysis Humanity's Last ExamKnowledge30.1%79 / 222
GPQA Diamond · Vals AIKnowledge84.5% ±1.850 / 135
AIME 2026Math95.3%13 / 34
Critical Physics TasksReasoning4.6%87 / 214
SWE-benchCoding76.4% ±1.937 / 87
LiveCodeBenchCoding81.4% ±1.162 / 142
Artificial Analysis Long Context ReasoningReasoning73.7%100 / 217
GDPval-AA normalizedAgentic30.8%62 / 133
SWE-bench ProCoding58.4%40 / 83
SWE-Bench verified · Epoch AICoding74.2% ±2.016 / 32
MCP AtlasAgentic71.8%29 / 52
Claw-EvalAgentic62.3%24 / 42
Humanity's Last ExamKnowledge31.0%44 / 77
Code MigrationCoding25.8% ±4.147 / 77
NL2RepoCoding42.7%23 / 36
Terminal-Bench 2.1 · Vals AIAgentic56.9% ±1.448 / 75
GPQA Diamond · BenchLMKnowledge86.2%54 / 84
Artificial Analysis SciCodeCoding44.8%74 / 114
Vibe Code Bench v1.1Coding31.5% ±4.681 / 123
Humanity's Last Exam with toolsAgentic52.3%24 / 36
SimpleQA VerifiedKnowledge34.0% ±1.561 / 91
Artificial Analysis Omniscience AccuracyKnowledge23.7%111 / 164
BrowseCompAgentic68.0%44 / 63
Artificial Analysis ITBench-AAAgentic40.3%17 / 24
Chess PuzzlesReasoning20.0% ±4.064 / 90
Harvard-MIT Mathematics Tournament February 2026Math82.6%20 / 27
CyberGymAgentic68.7%23 / 29
FrontierMath-Tiers-1-3-v2-PrivateMath24.9% ±2.673 / 90
OpenHarmony Bench v1.0Coding52.3%11 / 13
FrontierMath-Tier-4-2025-07-01-PublicMath0.0%3 / 36
FrontierMath-2025-02-28-PublicMath80.0% ±13.36 / 64
Finance Agent v1.1Finance57.7% ±2.810 / 51
LegalBenchLegal84.4% ±0.433 / 149
CorpFin v2Finance64.5% ±0.935 / 131
Public Benefits Bench v1Public benefits57.9% ±1.36 / 13
MedCodeHealthcare41.6% ±2.152 / 106
ProgramBenchCoding0.0%34 / 65
TaxEval v2Finance71.2% ±0.980 / 143
Public Benefits Bench v1.1Public benefits61.8% ±1.329 / 50
Legal Research BenchLegal27.9% ±3.161 / 90
Finance Agent (v2)Finance44.8% ±0.657 / 79
MedScribeHealthcare72.3% ±2.189 / 109
CaseLaw v2Legal51.6% ±1.147 / 53
MMAnswerBenchMath83.8%—
SWE-RebenchCoding62.7%—
ResearchClawBenchAgentic18.2%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYVals AINo licence statedmodels.devMIT

History

5 releases
Index56.9 → 56.0
23 Oct 202612 Nov 2026
Output / 1M$3.04 → $4.40
23 Oct 202612 Nov 2026

Questions

How much does GLM-5.1 cost?
$1.40 per 1M input tokens, $4.40 per 1M output tokens.
How good is GLM-5.1?
Its index is 56.0 out of 100. That is place 95 of 283 ranked models.
What is GLM-5.1 best at?
Agentic: 59.2 out of 100, place 77 of 218.
How fast is GLM-5.1?
102 output tokens per second, 0.5 s to the first token.
What is the context window of GLM-5.1?
205K tokens.
Is GLM-5.1 open source?
Yes. Its weights are published.
When was GLM-5.1 released?
April 2026.

Badge

GLM-5.1 on Inference360: place 95 of 283, index 56.0
HTML
<a href="https://inference360.ai/models/glm-5-1"><img src="https://inference360.ai/badges/glm-5-1.svg" alt="GLM-5.1 on Inference360: place 95 of 283, index 56.0" height="20"></a>
Markdown
[![GLM-5.1 on Inference360: place 95 of 283, index 56.0](https://inference360.ai/badges/glm-5-1.svg)](https://inference360.ai/models/glm-5-1)

Compare with

MiniMax M355.7Claude Opus 4.5 Thinking56.3Apodex 1.1 Mini55.7MiMo-V2.5-Pro56.3GLM-5.260.6

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeGLM-5.3-Flash63.1 · $0.078/1M

More from Z.AI

All models
GLM-5V-Turbo48.1GLM-5-Turbo51.9GLM-551.6GLM-5 (Reasoning)UnrankedGLM-5.260.6