Zhipu AI · mit

GLM-5.1

GLM-5.1 by Zhipu AI appears in 2 sources with Reasoning at 54.18. Best read for Coding, High intelligence, Low cost.

CreatorZhipu AI

Release date2026-04-07

Knowledge cutoffNot published

Context200K tokens

Input price$1.4/M tokens

Output price$4.4/M tokens

Modalitytext

CountryCN

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank5ranking · source_rank

Reasoning54.18reasoning · index_reasoning

Math46.76math · index_math

Coding42.99coding · index_code

Research29.67research · index_search

Vision39.03multimodal · index_vision

Tool calling28.89tool_calling · index_tool_calling

GPQA86.2 %reasoning · gpqa_score

Code Arena1,801.81 %coding · coding_arena_score

Humanity Last Exam52.3 %reasoning · hle_score

BrowseComp79.3 %research · browsecomp_score

Toolathlon40.7 %tool_calling · toolathlon_score

MCP Atlas71.8 %tool_calling · mcp_atlas_score

SWE-bench Pro58.4 %coding · swe_bench_pro_score

Context200,000 tokenscontext · context

Speed177.11 c/sperformance · throughput

Latency30,761.47 msperformance · latency

Input price1.4 $/Mpricing · input_price

Output price4.4 $/Mpricing · output_price

Parameters754,000,000,000 paramsmodel · params

Arena Rating1,469.34metric

Arena Rank11metric

Vote Count11,808metric

Evidence

Citations and source overlap

Arena AILM Arena text leaderboard (overall) mirrored in the public leaderboard dataset.Retrieved 2026-05-20 LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is GLM-5.1 verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.