llmrun Composite Score · Agents

The best open LLMs for agents, ranked on agentic and tool-use boards such as Terminal-Bench and tau-bench. The score shown is the overall llmrun Composite Score; the highlighted column is the agentic tool use skill score used for the order.

42 benchmarks4 sourcesUpdated 17 Oct 2026Reference scale 2026-Q4How the Composite Score is built

Agents score vs VRAM needed

This chart doesn't show proprietary models.

Open models Overall

General Coding Agents Math Science Reasoning

Models ranked by agents llmrun Composite Score
# (open / overall)	Model	llmrun Composite Score	Coding	Agents	Math	Science	Reasoning	Benchmarks	VRAM
1 / 19	GLM 5.3454 GB · 23 benchmarks		131	138	129	131	126	23	454 GB
2 / 28	GLM 5.3 Flash205 GB · 21 benchmarks		113	131	121	124	127	21	205 GB
3 / 35	GLM 5.2454 GB · 26 benchmarks		120	128	126	129	120	26	454 GB
4 / 42	GLM 5.1454 GB · 14 benchmarks		115	123	117	117	115	14	454 GB
5 / 43	Kimi K2.6618 GB · 20 benchmarks		119	122	122	122	119	20	618 GB
6 / 46	Kimi K2.7 Code618 GB · 19 benchmarks		115	119	121	118	118	19	618 GB
7 / 47	GLM 5454 GB · 15 benchmarks		110	118	–	–	109	15	454 GB
8 / 53	MiniMax M3258 GB · 19 benchmarks		113	113	106	118	107	19	258 GB
9 / 55	MiniMax M2.7140 GB · 7 benchmarks		103	112	–	112	–	7	140 GB
10 / 56	Kimi K2.5618 GB · 20 benchmarks		110	111	–	117	114	20	618 GB
11 / 57	MiniMax M2.5140 GB · 11 benchmarks		109	110	–	–	114	11	140 GB
12 / 59	Inkling575 GB · 22 benchmarks		104	109	112	116	121	22	575 GB
13 / 61	DeepSeek V3.2413 GB · 13 benchmarks		–	108	–	–	112	13	413 GB
14 / 62	Kimi K2 Thinking617 GB · 9 benchmarks		103	108	–	–	–	9	617 GB
15 / 63	GLM 4.7218 GB · 13 benchmarks		101	106	111	111	104	13	218 GB
16 / 64	GLM 4.5218 GB · 7 benchmarks		99	105	–	–	–	7	218 GB
17 / 66	Qwen3 Coder 480B A35B Instruct291 GB · 6 benchmarks	—	101	105	–	–	–	6	291 GB
18 / 67	MiniMax M2140 GB · 5 benchmarks		–	105	–	–	–	5	140 GB
19 / 68	GLM 4.6217 GB · 6 benchmarks		96	104	–	100	–	6	217 GB
20 / 72	Kimi K2 Instruct617 GB · 10 benchmarks		102	102	–	–	90	10	617 GB
21 / 78	Qwen2.5 Coder 32B Instruct24.3 GB · 3 benchmarks	—	–	84	–	–	–	3	24.3 GB
22 / 79	GPT OSS 120B70.8 GB · 20 benchmarks		98	82	–	103	97	20	70.8 GB

The range after each score is a 90% interval: given the boards a model has results on, its true score is very likely inside it. Short ranges mean many agreeing results; long ranges mean few or conflicting ones.

Why some models are missing. A model receives a score only if it has results on at least 4 benchmarks across at least 2 categories, with one being reasoning or coding. Skill scores require at least 2 benchmarks for that skill; otherwise they display as –.

VRAM is llmrun's estimate at Q4_K_M (or the smallest quantization we track) for the weights plus room for a 16K-token context. Longer contexts need more memory. Proprietary models can't run locally, so the VRAM filter hides them.

How the llmrun Composite Score is built · llmrun doesn't run these benchmarks; scores come from public sources.