Agentic · tools

JobBench leaderboard

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

As of 2026-10-07, the highest measured score on JobBench is 64.9% by Muse Spark 1.3. 150 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Opus 5	69.8%	estimated ± 3.4 pp, low confidence
2	GLM-5.3-Flash	66.2%	estimated ± 3.4 pp, low confidence
3	GPT-6 Astra	66.0%	estimated ± 4.9 pp, low confidence
4	Gemini 4 Argon	65.7%	estimated ± 4.9 pp, low confidence
5	Grok 4.7	65.4%	estimated ± 8.6 pp, low confidence
6	Claude Fable 5.1	65.2%	estimated ± 3.4 pp, low confidence
7	Claude Opus 5.5	65.2%	estimated ± 3.4 pp, low confidence
8	Claude Sonnet 5.5	65.2%	estimated ± 3.4 pp, low confidence
9	GPT-5.6 Sol	65.0%	estimated ± 4.9 pp, medium confidence
10	Muse Spark 1.3	64.9%	measured
11	GPT-6 Sol	64.8%	estimated ± 4.9 pp, medium confidence
12	Gemini 3.8 Flash	64.6%	estimated ± 4.9 pp, medium confidence
13	GPT-5.6 Terra	63.3%	estimated ± 4.9 pp, medium confidence
14	Gemini 3.7 Flash	62.9%	estimated ± 4.9 pp, medium confidence
15	GPT-5.6 Luna	62.5%	estimated ± 4.9 pp, medium confidence
16	MiMo-V2.6-Pro	62.0%	measured
17	Hy4 preview	61.7%	measured
18	MiMo-V2.6-Flash	61.2%	measured
19	Grok 4.6	59.5%	estimated ± 8.6 pp, medium confidence
20	Ling 3.1 Flash	59.5%	estimated ± 8.6 pp, medium confidence
21	Step 5 Preview	59.0%	measured
22	Claude Fable 5	58.9%	estimated ± 8.6 pp, medium confidence
23	DeepSeek V4 Pro 0813	58.3%	estimated ± 3.4 pp, low confidence
24	DeepSeek V4.1 Flash	58.2%	estimated ± 8.6 pp, medium confidence
25	GPT-6.1 Sol	56.7%	estimated ± 8.6 pp, medium confidence
26	GLM-5.3	56.1%	estimated ± 3.4 pp, low confidence
27	Qwen3.8 Max Preview	55.9%	estimated ± 8.2 pp, low confidence
28	Qwen3.8-Flash-Next	55.7%	measured
29	Muse Spark 1.1	54.7%	measured
30	Gemini 3.5 Flash	54.3%	estimated ± 6.8 pp, medium confidence
31	Fugu Cyber	53.9%	estimated ± 8.6 pp, low confidence
32	Gemini 3.8 Flash Cyber	53.5%	estimated ± 8.6 pp, low confidence
33	Claude Opus 4.8	53.4%	estimated ± 4.9 pp, medium confidence
34	Qwen3.8 Max	53.4%	measured
35	Kimi K3	52.9%	measured
36	Claude Mythos 5	52.4%	estimated ± 8.6 pp, low confidence
37	Ornith-1.5-397B	52.3%	estimated ± 3.4 pp, low confidence
38	Gemini 3.5 Flash Cyber	52.2%	estimated ± 8.6 pp, low confidence
39	Claude Mythos Preview	52.1%	estimated ± 8.6 pp, low confidence
40	GPT-5.5 Pro	51.8%	estimated ± 10.1 pp, low confidence
41	Holo3-35B-A3B	51.8%	estimated ± 9.0 pp, medium confidence
42	GPT-5.4 Pro	50.8%	estimated ± 10.1 pp, low confidence
43	Muse Spark 1.2	50.6%	estimated ± 8.6 pp, medium confidence
44	DeepSeek V4 Flash 0731	50.4%	estimated ± 3.4 pp, low confidence
45	Atria Dawn Preview	50.3%	measured
46	Claude Sonnet 5	49.9%	estimated ± 8.6 pp, medium confidence
47	Beam	49.1%	estimated ± 6.8 pp, medium confidence
48	Holo3-122B-A10B	49.0%	estimated ± 9.0 pp, medium confidence
49	GPT-6 Luna	48.1%	estimated ± 8.6 pp, medium confidence
50	Claude Opus 4.7	47.3%	estimated ± 4.9 pp, medium confidence
51	Mistral Large 4	47.3%	estimated ± 8.6 pp, medium confidence
52	GLM-5.2	47.1%	estimated ± 6.8 pp, medium confidence
53	Qwen3.7 Max	46.7%	estimated ± 6.8 pp, medium confidence
54	Kimi K2.7 Code	46.4%	estimated ± 6.8 pp, medium confidence
55	Claude Opus 4.7 (Adaptive)	45.9%	measured
56	Muse Glimmer 30B	45.9%	estimated ± 6.8 pp, medium confidence
57	K-EXAONE 2.0	45.9%	estimated ± 8.4 pp, low confidence
58	Ornith-1.0-397B	45.5%	estimated ± 8.4 pp, low confidence
59	Grok 4.5	45.2%	estimated ± 8.6 pp, medium confidence
60	UI-Mate-27B	45.0%	estimated ± 9.0 pp, medium confidence
61	Inkling	44.5%	estimated ± 6.8 pp, medium confidence
62	GPT-5.5	42.7%	measured
63	GLM-5.1	42.4%	estimated ± 6.8 pp, medium confidence
64	Qwen3.6-27B	42.3%	estimated ± 8.4 pp, low confidence
65	Claude Opus 4.6 (Adaptive)	41.4%	estimated ± 8.2 pp, low confidence
66	Ornith-1.0-35B	40.1%	estimated ± 8.4 pp, low confidence
67	Gemini 3.1 Pro	39.9%	estimated ± 8.2 pp, low confidence
68	GPT-5.4	38.9%	measured
69	Gemini 3.6 Flash	38.7%	estimated ± 8.6 pp, medium confidence
70	Apodex 1.1	38.7%	estimated ± 8.2 pp, low confidence
71	Apodex 1.1 Mini	38.7%	estimated ± 8.2 pp, low confidence
72	Claude Sonnet 4.6	36.9%	measured
73	Claude Opus 4.6	36.7%	measured
74	Step 3.7 Flash	36.0%	estimated ± 7.3 pp, low confidence
75	Hy3 Preview	34.4%	estimated ± 8.6 pp, medium confidence
76	GPT-5.2	34.3%	measured
77	MiniMax M2.7	34.1%	estimated ± 7.3 pp, low confidence
78	Agents-A1	33.9%	estimated ± 10.1 pp, low confidence
79	GPT-5.3 Codex	33.7%	measured
80	Muse Spark	33.6%	estimated ± 8.4 pp, low confidence
81	Solar Pro 4	33.5%	estimated ± 6.8 pp, medium confidence
82	Qwen3.8-27B	33.4%	measured
83	UI-Mate-9B	32.8%	estimated ± 9.0 pp, medium confidence
84	Ornith-1.0-9B	32.8%	estimated ± 8.4 pp, low confidence
85	Qwen3.5-27B	32.8%	estimated ± 9.0 pp, medium confidence
86	Qwen3.5-35B-A3B	32.8%	estimated ± 9.0 pp, medium confidence
87	LFM2.5-2.6B	32.5%	estimated ± 8.4 pp, low confidence
88	Quasar 438B	32.4%	estimated ± 8.6 pp, medium confidence
89	Claude Opus 4.5	32.3%	measured
90	MiMo-V2.5	31.8%	estimated ± 8.4 pp, low confidence
91	Ling 3.0 Flash VL	31.1%	estimated ± 8.6 pp, medium confidence
92	Solar Open 2	31.0%	estimated ± 6.8 pp, medium confidence
93	Nemotron 3 Ultra	31.0%	estimated ± 8.6 pp, medium confidence
94	GPT-5.4 mini	30.7%	estimated ± 6.8 pp, medium confidence
95	GPT-5.4 nano	29.5%	estimated ± 6.8 pp, medium confidence
96	Kimi K2.6	27.7%	estimated ± 4.9 pp, low confidence
97	MiniMax M3	27.7%	estimated ± 4.9 pp, low confidence
98	Claude Sonnet 4.5	27.7%	measured
99	GPT-5.1-Codex	26.2%	measured
100	GPT-5.2-Codex	26.0%	measured
101	MiMo-V2-Pro	25.7%	estimated ± 8.4 pp, low confidence
102	Hy3	25.1%	estimated ± 8.6 pp, medium confidence
103	LLaDA2.2-mini	24.8%	estimated ± 8.4 pp, low confidence
104	Agents-A1-4B	24.7%	estimated ± 10.1 pp, low confidence
105	GLM-5-Turbo	22.8%	estimated ± 8.4 pp, low confidence
106	LLaDA2.2-flash	22.7%	estimated ± 6.8 pp, medium confidence
107	LongCat-Flash-Lite-Sparse	22.3%	estimated ± 6.8 pp, medium confidence
108	GLM-4.7	22.0%	estimated ± 8.6 pp, medium confidence
109	Claude 4.1 Opus	21.9%	measured
110	Gemini 3.5 Flash-Lite	20.1%	estimated ± 8.6 pp, medium confidence
111	GLM-5V-Turbo	20.0%	estimated ± 8.4 pp, low confidence
112	dots3-note Preview	19.9%	estimated ± 3.4 pp, low confidence
113	Qwen3.7 Plus	19.9%	estimated ± 4.9 pp, low confidence
114	A.X K2	18.8%	estimated ± 8.6 pp, medium confidence
115	Qwen3.6 Plus	18.5%	estimated ± 5.1 pp, low confidence
116	Qwen3.5 Plus	18.5%	measured
117	Claude 4 Sonnet	18.4%	measured
118	Inkling-Small	17.9%	estimated ± 3.4 pp, low confidence
119	Ling 3.0 Flash FP8	17.8%	estimated ± 8.6 pp, medium confidence
120	Qwen3.5 397B	17.2%	estimated ± 5.1 pp, low confidence
121	Grok 4.3	16.8%	estimated ± 8.2 pp, low confidence
122	Claude Haiku 4.5	16.0%	measured
123	Ling 3.0 Flash	15.4%	estimated ± 5.1 pp, low confidence
124	Gemini 3 Flash	11.4%	measured
125	Gemini 3 Pro	11.4%	measured
126	Laguna S 2.1	11.3%	estimated ± 3.4 pp, low confidence
127	GPT-5.1	10.4%	estimated ± 8.6 pp, low confidence
128	Ornith-1.5-35B-A3B	10.1%	estimated ± 3.4 pp, low confidence
129	Qwen3.5-122B-A10B	9.6%	estimated ± 8.6 pp, low confidence
130	MiMo-V2-Omni	9.2%	estimated ± 8.4 pp, low confidence
131	Kimi K2.5	8.7%	measured
132	GPT-5 (high)	8.5%	measured
133	Kimi K2.5 (Reasoning)	8.3%	estimated ± 8.2 pp, low confidence
134	Mistral Medium 3.5 128B	6.4%	estimated ± 8.6 pp, low confidence
135	DeepSeek V3.2	5.0%	estimated ± 8.4 pp, low confidence
136	Ornith-1.5-9B	4.0%	estimated ± 3.4 pp, low confidence
137	MiniCPM5-2B	3.0%	estimated ± 8.6 pp, low confidence
138	Claude Haiku 5.5	0.7%	estimated ± 9.1 pp, low confidence
139	Celeris-1	0.0%	estimated ± 8.6 pp, low confidence
140	Command A+	0.0%	estimated ± 8.6 pp, low confidence
141	DeepSeek V3	0.0%	estimated ± 8.6 pp, low confidence
142	DeepSeek V3 0324	0.0%	estimated ± 8.6 pp, low confidence
143	Gemini 2.5 Pro	0.0%	estimated ± 8.6 pp, low confidence
144	Gemma 3 27B	0.0%	estimated ± 8.6 pp, low confidence
145	Gemma 4 12B	0.0%	estimated ± 8.6 pp, low confidence
146	Gemma 4 26B A4B	0.0%	estimated ± 8.6 pp, low confidence
147	Gemma 4 31B	0.0%	estimated ± 8.6 pp, low confidence
148	Gemma 4 E2B	0.0%	estimated ± 8.6 pp, low confidence
149	Gemma 4 E4B	0.0%	estimated ± 8.6 pp, low confidence
150	GLM-5	0.0%	estimated ± 5.1 pp, low confidence
151	GPT-4.1 mini	0.0%	estimated ± 8.6 pp, low confidence
152	GPT-4.1 nano	0.0%	estimated ± 8.6 pp, low confidence
153	GPT-4o	0.0%	estimated ± 8.6 pp, low confidence
154	GPT-4o mini	0.0%	estimated ± 8.6 pp, low confidence
155	GPT-OSS 120B	0.0%	estimated ± 8.2 pp, low confidence
156	GPT-OSS 20B	0.0%	estimated ± 8.2 pp, low confidence
157	Granite 4.2 30B	0.0%	estimated ± 8.6 pp, low confidence
158	Granite 4.2 3B	0.0%	estimated ± 8.6 pp, low confidence
159	Granite 4.2 8B	0.0%	estimated ± 8.6 pp, low confidence
160	K-Exaone	0.0%	estimated ± 8.6 pp, low confidence
161	Ling 2.6 Flash	0.0%	estimated ± 8.6 pp, low confidence
162	Ling 3.0 Tiny	0.0%	estimated ± 8.6 pp, low confidence
163	Llama 4 Maverick	0.0%	estimated ± 8.6 pp, low confidence
164	Llama 4 Scout	0.0%	estimated ± 8.6 pp, low confidence
165	Mercury 2.5	0.0%	estimated ± 8.6 pp, low confidence
166	MiMo-V2.5-Pro	0.0%	estimated ± 8.2 pp, low confidence
167	MiMo-V2-Flash	0.0%	estimated ± 8.6 pp, low confidence
168	Mistral Large 3	0.0%	estimated ± 8.6 pp, low confidence
169	Mistral Small 4	0.0%	estimated ± 8.6 pp, low confidence
170	Mistral Small 4 (Reasoning)	0.0%	estimated ± 8.6 pp, low confidence
171	Nemotron 3.5 Lightning 30B A3B NVFP4	0.0%	estimated ± 8.6 pp, low confidence
172	Nemotron 3 Nano 30B	0.0%	estimated ± 8.6 pp, low confidence
173	Nemotron 3 Nano Omni 30B A3B	0.0%	estimated ± 8.6 pp, low confidence
174	Nemotron 3 Super 100B	0.0%	estimated ± 8.2 pp, low confidence
175	North Mini Code	0.0%	estimated ± 8.6 pp, low confidence
176	Qwen3.6-35B-A3B	0.0%	estimated ± 5.1 pp, low confidence
177	Solar Pro 3	0.0%	estimated ± 8.6 pp, low confidence
178	Trinity-Large-Preview	0.0%	estimated ± 8.6 pp, low confidence
179	Trinity-Large-Thinking	0.0%	estimated ± 8.6 pp, low confidence
180	Ultravox v0.6 Llama 3.3 70B	0.0%	estimated ± 8.6 pp, low confidence

JobBench leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General