Pass IndexThe State of AISign in

AgentRL

AgentRL is a model by Z.ai for reasoning and agentic, it stands 1st of 25 on AgentBench Leaderboard, and 4 sellers in the catalogue publish a price for it.

text → text · made by Z.ai

$0.8$0.8per Mtok in / out
Together AI · 4 sellers

Sold by 4 ways

SellerLaneRate
Together AI aggregatorstandard$0.8per Mtok in$0.8per Mtok outraw.githubusercontent.com · read 2026-09-16
Fireworks AI aggregatorstandard$0.9per Mtok in$0.9per Mtok outraw.githubusercontent.com · read 2026-09-16
ElectronHub aggregatorstandard$1per Mtok in$1per Mtok outapi.electronhub.ai · read 2026-09-16
PPIO aggregatorstandard$1.26per Mtok in$1.26per Mtok outapi.ppinfra.com · read 2026-09-10

Measured 7 standings

PlaceBoardMetricScore
1stof 25AgentBench Leaderboard (new)Success Rate (pass@1) AVG70.4
29thof 217Mmlu — Epoch AIEM0.799
57thof 108Math level 5 — Epoch AImean_score56.071
199thof 222Chess puzzles — Epoch AImean_score0
246thof 290OTIS Mock AIME 2024-2025mean_score7.361
249thof 313GPQA Diamondmean_score46.086
538thof 656Epoch capabilities index — Epoch AIECI Score128.53

About

AgentRL — a text model from Z.ai, sold by 4 companies from $0.8 in and $0.8 out per million tokens, placed 1st of 25 on AgentBench Leaderboard (new).

It takes text and returns text, with a context window of 131,072 tokens. It was published in September 2024, trained on material up to April 2024. Its sellers say it can call a tool. The catalogue files it under reasoning and agents. Four companies sell it. The cheapest is $0.8 in and $0.8 out per million tokens at Together AI. It has been measured on 7 boards, and stands best at 1st of 25 on AgentBench Leaderboard (new).

Every current figure

Maker
Z.ai
Register
model
Takes
text
Returns
text
Context
131,072 tokens
Published
September 2024
Knowledge to
April 2024
Sellers
4
Maker's own price
not read
Price
$0.8 in and $0.8 out per million tokens — Together AI
Boards
7
Best place
1st of 25 — AgentBench Leaderboard (new)

Known as 6 names

AgentRL w/ GLM-4-9B-0414AgentRL w/ Qwen2.5-14B-InstructAgentRL w/ Qwen2.5-32B-InstructAgentRL w/ Qwen2.5-3B-InstructQwen2.5-14B-InstructQwen2.5-32B-Instruct