Agentic · tools

Toolathlon leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

As of 2026-10-07, the highest measured score on Toolathlon is 75.6% by Muse Spark 1.1. 154 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Mercury 2.5	90.3%	estimated ± 6.0 pp, low confidence
2	Mistral Medium 3.5 128B	83.3%	estimated ± 6.0 pp, low confidence
3	Muse Spark 1.1	75.6%	measured
4	Claude Fable 5.1	67.2%	estimated ± 1.7 pp, low confidence
5	GPT-6 Astra	66.4%	estimated ± 1.7 pp, low confidence
6	Holo3-35B-A3B	63.8%	estimated ± 7.6 pp, medium confidence
7	Claude Opus 4.8	59.9%	measured
8	Claude Opus 5	59.3%	estimated ± 1.7 pp, low confidence
9	Claude Fable 5	58.9%	estimated ± 1.7 pp, low confidence
10	Holo3-122B-A10B	58.5%	estimated ± 7.6 pp, medium confidence
11	GPT-5.6 Sol	58.0%	measured
12	Atria Dawn Preview	57.3%	estimated ± 3.5 pp, medium confidence
13	Gemini 3.8 Flash	56.7%	estimated ± 1.7 pp, medium confidence
14	GPT-5.5 Pro	56.6%	estimated ± 3.5 pp, high confidence
15	Gemini 3.5 Flash	56.5%	measured
16	GPT-5.4 Pro	56.4%	estimated ± 3.5 pp, high confidence
17	Step 5 Preview	56.2%	estimated ± 3.5 pp, high confidence
18	UI-Mate-27B	56.0%	estimated ± 7.6 pp, medium confidence
19	Claude Mythos 5	55.9%	estimated ± 3.5 pp, high confidence
20	Claude Opus 5.5	55.9%	estimated ± 6.3 pp, low confidence
21	Claude Sonnet 5.5	55.9%	estimated ± 6.3 pp, low confidence
22	DeepSeek V4.1 Flash	55.9%	estimated ± 6.3 pp, medium confidence
23	Gemini 4 Argon	55.9%	estimated ± 6.3 pp, low confidence
24	GLM-5.3	55.9%	estimated ± 6.3 pp, low confidence
25	Grok 4.7	55.9%	estimated ± 6.3 pp, low confidence
26	Ling 3.1 Flash	55.9%	estimated ± 6.3 pp, low confidence
27	MiMo-V2.6-Flash	55.9%	estimated ± 6.3 pp, medium confidence
28	MiMo-V2.6-Pro	55.9%	estimated ± 6.3 pp, low confidence
29	Qwen3.8-Flash-Next	55.9%	estimated ± 6.3 pp, low confidence
30	Qwen3.8 Max Preview	55.9%	estimated ± 6.3 pp, low confidence
31	GPT-6.1 Sol	55.9%	estimated ± 6.3 pp, medium confidence
32	GPT-6 Luna	55.9%	estimated ± 6.3 pp, medium confidence
33	GPT-6 Sol	55.9%	estimated ± 6.3 pp, medium confidence
34	Mistral Large 4	55.9%	estimated ± 6.3 pp, medium confidence
35	Muse Spark 1.2	55.9%	estimated ± 6.3 pp, medium confidence
36	Qwen3.8-27B	55.9%	estimated ± 6.3 pp, medium confidence
37	Grok 4.5	55.9%	estimated ± 6.3 pp, medium confidence
38	Gemini 3.6 Flash	55.8%	estimated ± 6.3 pp, medium confidence
39	Hy3 Preview	55.8%	estimated ± 6.3 pp, medium confidence
40	Apodex 1.1	55.8%	estimated ± 6.3 pp, medium confidence
41	Apodex 1.1 Mini	55.8%	estimated ± 6.3 pp, medium confidence
42	Quasar 438B	55.7%	estimated ± 6.3 pp, medium confidence
43	Ling 3.0 Flash VL	55.7%	estimated ± 6.3 pp, medium confidence
44	GPT-5.5	55.6%	measured
45	Muse Spark 1.3	55.6%	estimated ± 1.7 pp, medium confidence
46	Ornith-1.5-397B	55.4%	estimated ± 3.5 pp, high confidence
47	Kimi K3	55.4%	estimated ± 1.7 pp, medium confidence
48	Qwen3.7 Max	55.4%	estimated ± 6.3 pp, medium confidence
49	Claude Mythos Preview	54.9%	estimated ± 9.0 pp, low confidence
50	Claude Sonnet 5	54.9%	estimated ± 1.7 pp, medium confidence
51	Gemini 3.7 Flash	54.9%	estimated ± 1.7 pp, medium confidence
52	GPT-5.4	54.6%	measured
53	Grok 4.6	54.3%	estimated ± 1.7 pp, medium confidence
54	MiniMax M3	54.0%	estimated ± 3.5 pp, high confidence
55	Grok 4.3	54.0%	estimated ± 6.3 pp, medium confidence
56	dots3-note Preview	53.9%	estimated ± 3.5 pp, high confidence
57	GPT-5.6 Luna	53.4%	measured
58	GPT-5.6 Terra	53.1%	measured
59	Claude Opus 4.7	53.0%	estimated ± 1.7 pp, medium confidence
60	Hy3	52.8%	estimated ± 6.3 pp, medium confidence
61	Hy4 preview	52.5%	estimated ± 5.2 pp, low confidence
62	Claude Opus 4.6	52.5%	estimated ± 1.7 pp, low confidence
63	MiMo-V2.5	52.3%	estimated ± 6.9 pp, medium confidence
64	Claude Sonnet 4.6	51.9%	estimated ± 1.7 pp, low confidence
65	DeepSeek V4 Pro 0813	51.8%	measured
66	Claude Opus 4.7 (Adaptive)	51.7%	estimated ± 3.5 pp, high confidence
67	Inkling-Small	50.4%	estimated ± 3.5 pp, high confidence
68	Beam	50.4%	estimated ± 3.5 pp, high confidence
69	Claude Opus 4.6 (Adaptive)	50.3%	estimated ± 7.0 pp, low confidence
70	Inkling	50.2%	estimated ± 3.5 pp, high confidence
71	Gemini 3 Pro	50.1%	estimated ± 7.2 pp, medium confidence
72	Kimi K2.7 Code	50.1%	estimated ± 6.3 pp, medium confidence
73	Kimi K2.6	50.0%	measured
74	MiMo-V2-Pro	49.7%	estimated ± 6.9 pp, medium confidence
75	Step 3.7 Flash	49.5%	measured
76	Qwen3.8 Max	49.4%	estimated ± 5.2 pp, low confidence
77	Agents-A1	48.9%	estimated ± 3.5 pp, high confidence
78	GLM-5.2	48.2%	measured
79	DeepSeek V4 Flash 0731	47.8%	measured
80	GPT-5.3 Codex	46.6%	estimated ± 7.2 pp, medium confidence
81	MiniMax M2.7	46.3%	measured
82	Gemini 3 Flash	46.1%	estimated ± 7.2 pp, medium confidence
83	Ling 3.0 Flash	46.0%	estimated ± 3.5 pp, high confidence
84	MiMo-V2.5-Pro	45.6%	estimated ± 6.0 pp, low confidence
85	UI-Mate-9B	43.8%	estimated ± 7.6 pp, medium confidence
86	Claude Opus 4.5	43.5%	measured
87	Muse Spark	43.3%	estimated ± 6.3 pp, medium confidence
88	GPT-5.2-Codex	43.0%	estimated ± 7.2 pp, medium confidence
89	GPT-5.4 mini	42.9%	measured
90	GPT-5.1-Codex	41.6%	estimated ± 7.2 pp, medium confidence
91	GLM-5.1	41.4%	estimated ± 3.5 pp, high confidence
92	Grok Build 0.1	41.2%	estimated ± 7.2 pp, medium confidence
93	Ornith-1.5-35B-A3B	40.9%	estimated ± 3.5 pp, high confidence
94	Claude Sonnet 4.5	40.8%	estimated ± 7.2 pp, medium confidence
95	Qwen3.6-27B	40.4%	estimated ± 6.3 pp, medium confidence
96	Solar Open 2	40.3%	estimated ± 7.5 pp, medium confidence
97	Grok 4.1 Fast	40.0%	estimated ± 7.2 pp, medium confidence
98	Gemini 3.5 Flash-Lite	40.0%	estimated ± 6.3 pp, medium confidence
99	Agents-A1-4B	39.9%	estimated ± 3.5 pp, high confidence
100	Qwen3.6 Plus	39.8%	measured
101	GPT-5.2	38.6%	estimated ± 3.5 pp, high confidence
102	GLM-5	38.0%	measured
103	Qwen3 Max	37.5%	estimated ± 7.2 pp, medium confidence
104	LLaDA2.2-flash	37.2%	estimated ± 7.5 pp, medium confidence
105	Grok 4	36.6%	estimated ± 7.2 pp, medium confidence
106	Qwen3.5 397B	36.3%	measured
107	Qwen3.5-122B-A10B	35.9%	estimated ± 3.5 pp, high confidence
108	A.X K2	35.6%	estimated ± 6.3 pp, medium confidence
109	GPT-5.4 nano	35.5%	measured
110	Claude 4 Sonnet	34.7%	estimated ± 7.2 pp, low confidence
111	Gemini 3.1 Flash-Lite	33.8%	estimated ± 7.2 pp, low confidence
112	Grok 4.20	33.7%	estimated ± 7.2 pp, low confidence
113	Claude 4.1 Opus	33.1%	estimated ± 7.4 pp, low confidence
114	Ling 3.0 Flash FP8	33.1%	estimated ± 6.3 pp, medium confidence
115	Pokee-Isaac 28B	31.8%	estimated ± 6.0 pp, low confidence
116	Qwen3.5-27B	31.8%	estimated ± 3.5 pp, high confidence
117	Qwen3.5-35B-A3B	31.8%	estimated ± 3.5 pp, high confidence
118	Kimi K2.5 (Reasoning)	31.2%	estimated ± 3.5 pp, high confidence
119	Qwen3.5 Plus	30.9%	estimated ± 7.4 pp, low confidence
120	GPT-5 (high)	30.0%	estimated ± 6.3 pp, medium confidence
121	Claude Haiku 4.5	29.7%	estimated ± 7.4 pp, low confidence
122	GLM-5V-Turbo	28.0%	estimated ± 7.2 pp, low confidence
123	Kimi K2.5	27.8%	measured
124	GPT-5.1	27.2%	estimated ± 6.3 pp, low confidence
125	DeepSeek V3.2	27.0%	estimated ± 7.2 pp, low confidence
126	Gemini 3.1 Pro	27.0%	estimated ± 6.3 pp, low confidence
127	Muse Glimmer 30B	27.0%	estimated ± 6.3 pp, low confidence
128	Qwen3.7 Plus	27.0%	estimated ± 6.3 pp, low confidence
129	MiniCPM5-2B	26.9%	estimated ± 6.3 pp, low confidence
130	MiMo-V2-Flash	26.9%	estimated ± 6.3 pp, low confidence
131	Gemma 4 31B	26.9%	estimated ± 6.3 pp, low confidence
132	GPT-OSS 120B	26.9%	estimated ± 6.3 pp, low confidence
133	Gemma 4 26B A4B	26.9%	estimated ± 6.3 pp, low confidence
134	Ling 3.0 Tiny	26.9%	estimated ± 6.3 pp, low confidence
135	Celeris-1	26.9%	estimated ± 6.3 pp, low confidence
136	Command A+	26.9%	estimated ± 6.3 pp, low confidence
137	DeepSeek V3	26.9%	estimated ± 6.3 pp, low confidence
138	DeepSeek V3 0324	26.9%	estimated ± 6.3 pp, low confidence
139	Gemini 2.5 Pro	26.9%	estimated ± 6.3 pp, low confidence
140	Gemma 3 27B	26.9%	estimated ± 6.3 pp, low confidence
141	Gemma 4 12B	26.9%	estimated ± 6.3 pp, low confidence
142	Gemma 4 E2B	26.9%	estimated ± 6.3 pp, low confidence
143	Gemma 4 E4B	26.9%	estimated ± 6.3 pp, low confidence
144	GPT-4.1 mini	26.9%	estimated ± 6.3 pp, low confidence
145	GPT-4.1 nano	26.9%	estimated ± 6.3 pp, low confidence
146	GPT-4o	26.9%	estimated ± 6.3 pp, low confidence
147	GPT-4o mini	26.9%	estimated ± 6.3 pp, low confidence
148	GPT-OSS 20B	26.9%	estimated ± 6.3 pp, low confidence
149	K-Exaone	26.9%	estimated ± 6.3 pp, low confidence
150	Ling 2.6 Flash	26.9%	estimated ± 6.3 pp, low confidence
151	Llama 4 Maverick	26.9%	estimated ± 6.3 pp, low confidence
152	Llama 4 Scout	26.9%	estimated ± 6.3 pp, low confidence
153	Mistral Large 3	26.9%	estimated ± 6.3 pp, low confidence
154	Mistral Small 4	26.9%	estimated ± 6.3 pp, low confidence
155	Mistral Small 4 (Reasoning)	26.9%	estimated ± 6.3 pp, low confidence
156	Nemotron 3 Nano 30B	26.9%	estimated ± 6.3 pp, low confidence
157	Nemotron 3 Nano Omni 30B A3B	26.9%	estimated ± 6.3 pp, low confidence
158	Nemotron 3 Super 100B	26.9%	estimated ± 6.3 pp, low confidence
159	North Mini Code	26.9%	estimated ± 6.3 pp, low confidence
160	Solar Pro 3	26.9%	estimated ± 6.3 pp, low confidence
161	Trinity-Large-Preview	26.9%	estimated ± 6.3 pp, low confidence
162	Trinity-Large-Thinking	26.9%	estimated ± 6.3 pp, low confidence
163	Ultravox v0.6 Llama 3.3 70B	26.9%	estimated ± 6.3 pp, low confidence
164	Qwen3.6-35B-A3B	26.9%	measured
165	Ornith-1.5-9B	24.6%	estimated ± 3.5 pp, medium confidence
166	Granite 4.2 30B	24.5%	estimated ± 6.0 pp, low confidence
167	GPT-4.1	23.9%	estimated ± 7.2 pp, low confidence
168	Granite 4.2 8B	18.8%	estimated ± 6.0 pp, low confidence
169	GLM-4.7	17.9%	estimated ± 3.5 pp, medium confidence
170	Grok 4.1	17.3%	estimated ± 6.9 pp, low confidence
171	Solar Pro 4	13.9%	estimated ± 3.5 pp, medium confidence
172	LongCat-Flash-Lite-Sparse	13.2%	estimated ± 3.5 pp, medium confidence
173	Nemotron 3 Ultra	8.4%	estimated ± 3.5 pp, medium confidence
174	Granite 4.2 3B	8.3%	estimated ± 6.0 pp, low confidence
175	LFM2.5-2.6B	4.5%	estimated ± 6.0 pp, low confidence
176	Nemotron 3.5 Lightning 30B A3B NVFP4	3.0%	estimated ± 3.5 pp, medium confidence

Toolathlon leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General