Knowledge & reasoning

C-Eval leaderboard

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

As of 2026-10-07, the highest measured score on C-Eval is 93.3% by Qwen3.6 Plus. 225 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Opus 4.6	100.0%	estimated ± 0.7 pp, low confidence
2	Claude Sonnet 4.6	100.0%	estimated ± 0.7 pp, low confidence
3	Gemini 3 Pro Deep Think	100.0%	estimated ± 1.4 pp, low confidence
4	Claude Opus 5.5	97.1%	estimated ± 1.0 pp, low confidence
5	Claude Sonnet 5.5	97.0%	estimated ± 1.0 pp, low confidence
6	Gemini 4 Argon	96.8%	estimated ± 1.0 pp, low confidence
7	GPT-6.1 Sol	96.7%	estimated ± 1.0 pp, low confidence
8	Claude Fable 5	96.6%	estimated ± 1.0 pp, low confidence
9	Muse Spark 1.3	96.4%	estimated ± 1.0 pp, low confidence
10	GPT-6 Sol	96.4%	estimated ± 1.0 pp, low confidence
11	GPT-6 Astra	96.4%	estimated ± 0.9 pp, low confidence
12	Grok 4.7	96.3%	estimated ± 1.0 pp, low confidence
13	MiMo-V2.6-Pro	96.3%	estimated ± 1.0 pp, low confidence
14	Qwen3.8 Max Preview	96.2%	estimated ± 1.0 pp, low confidence
15	GLM-5.3	96.1%	estimated ± 1.0 pp, low confidence
16	Sakana Fugu	96.1%	estimated ± 0.9 pp, low confidence
17	Sakana Fugu-Ultra	96.1%	estimated ± 0.9 pp, low confidence
18	Grok 4.6	96.1%	estimated ± 1.0 pp, low confidence
19	Claude Haiku 5.5	96.0%	estimated ± 1.0 pp, low confidence
20	GLM-5.3-Flash	95.8%	estimated ± 1.0 pp, low confidence
21	Ling 3.1 Flash	95.8%	estimated ± 1.0 pp, low confidence
22	Gemini 3.8 Flash	95.8%	estimated ± 1.0 pp, low confidence
23	GPT-5.6 Sol	95.6%	estimated ± 0.9 pp, low confidence
24	Muse Spark 1.2	95.6%	estimated ± 1.0 pp, low confidence
25	Gemini 3.7 Flash	95.5%	estimated ± 1.0 pp, low confidence
26	Grok 4.5	95.5%	estimated ± 1.0 pp, low confidence
27	Mistral Large 4	95.5%	estimated ± 1.0 pp, low confidence
28	GPT-6 Luna	95.4%	estimated ± 1.0 pp, low confidence
29	MiMo-V2.6-Flash	95.4%	estimated ± 1.0 pp, low confidence
30	Gemini 3.6 Flash	94.8%	estimated ± 1.0 pp, low confidence
31	GPT-5.6 Terra	94.7%	estimated ± 0.9 pp, low confidence
32	GPT-5.3 Codex	94.6%	estimated ± 1.0 pp, low confidence
33	Claude Opus 4.6 (Adaptive)	94.5%	estimated ± 1.0 pp, low confidence
34	GPT-5.2	94.5%	estimated ± 0.9 pp, low confidence
35	GPT-5.6 Luna	94.4%	estimated ± 0.9 pp, low confidence
36	Claude Opus 4.7	94.3%	estimated ± 1.0 pp, low confidence
37	Gemini 3.1 Pro	94.1%	estimated ± 1.0 pp, low confidence
38	Qwen 3.6 Max (preview)	94.0%	estimated ± 0.7 pp, low confidence
39	MiniMax M3	94.0%	estimated ± 1.0 pp, low confidence
40	Claude Opus 4.5 Thinking	93.9%	estimated ± 1.0 pp, low confidence
41	Qwen3.7 Max	93.9%	estimated ± 0.7 pp, low confidence
42	MiMo-V2-Pro	93.8%	estimated ± 1.0 pp, low confidence
43	GPT-5.2-Codex	93.8%	estimated ± 1.0 pp, low confidence
44	Solar Pro 4	93.7%	estimated ± 1.0 pp, low confidence
45	Gemini 3 Pro	93.7%	estimated ± 1.0 pp, low confidence
46	Quasar 438B	93.4%	estimated ± 1.0 pp, medium confidence
47	GLM-5-Turbo	93.4%	estimated ± 1.0 pp, medium confidence
48	Apodex 1.1 Mini	93.3%	estimated ± 1.0 pp, medium confidence
49	Qwen3.6 Plus	93.3%	measured
50	Claude Fable 5.1	93.3%	estimated ± 0.8 pp, low confidence
51	Claude Opus 5	93.3%	estimated ± 0.8 pp, low confidence
52	Claude Mythos 5	93.3%	estimated ± 0.8 pp, low confidence
53	Muse Spark 1.1	93.3%	estimated ± 0.8 pp, low confidence
54	GPT-5.4 Pro	93.3%	estimated ± 0.8 pp, low confidence
55	Claude Opus 4.8	93.3%	estimated ± 0.8 pp, low confidence
56	Claude Sonnet 5	93.3%	estimated ± 0.8 pp, low confidence
57	GPT-5.5 Pro	93.3%	estimated ± 0.8 pp, low confidence
58	Apodex 1.1	93.3%	estimated ± 0.8 pp, low confidence
59	Kimi K3	93.3%	estimated ± 0.8 pp, low confidence
60	Hy4 preview	93.3%	estimated ± 0.8 pp, low confidence
61	Claude Opus 4.7 (Adaptive)	93.3%	estimated ± 0.8 pp, low confidence
62	GLM-5.2	93.3%	estimated ± 0.8 pp, low confidence
63	dots3-note Preview	93.3%	estimated ± 0.8 pp, low confidence
64	GLM-5.1	93.3%	estimated ± 0.8 pp, low confidence
65	GPT-5.5	93.3%	estimated ± 0.8 pp, low confidence
66	GPT-5.4	93.3%	estimated ± 0.8 pp, low confidence
67	Muse Spark	93.3%	estimated ± 0.8 pp, low confidence
68	MiMo-V2.5-Pro	93.3%	estimated ± 0.8 pp, low confidence
69	Inkling-Small	93.3%	estimated ± 0.8 pp, low confidence
70	Agents-A1	93.3%	estimated ± 0.8 pp, low confidence
71	Step 5 Preview	93.3%	estimated ± 0.8 pp, low confidence
72	Inkling	93.3%	estimated ± 0.8 pp, low confidence
73	Ornith-1.5-397B	93.3%	estimated ± 0.8 pp, low confidence
74	Qwen3.8 Max	93.2%	estimated ± 0.8 pp, low confidence
75	DeepSeek V4 Pro 0813	93.2%	estimated ± 0.8 pp, low confidence
76	GPT-5.4 mini	93.2%	estimated ± 0.8 pp, low confidence
77	Gemini 3.5 Flash	93.2%	estimated ± 0.8 pp, low confidence
78	GPT-5.4 nano	93.2%	estimated ± 0.8 pp, low confidence
79	DeepSeek V4.1 Flash	93.2%	estimated ± 0.8 pp, low confidence
80	Qwen3.8-Omni-Flash	93.2%	estimated ± 0.8 pp, low confidence
81	Kimi K2.7 Code	93.1%	estimated ± 1.0 pp, medium confidence
82	Qwen3.8-Flash-Next	93.1%	estimated ± 0.8 pp, low confidence
83	Grok 4.3	93.1%	estimated ± 0.8 pp, low confidence
84	DeepSeek V4 Flash 0731	93.1%	estimated ± 0.8 pp, low confidence
85	Kimi K2.6	93.1%	estimated ± 0.8 pp, low confidence
86	Interfaze Beta	93.1%	estimated ± 0.9 pp, medium confidence
87	Qwen3.7 Plus	93.1%	estimated ± 0.7 pp, medium confidence
88	Hy3	93.0%	estimated ± 1.0 pp, medium confidence
89	Qwen3.5 397B	93.0%	measured
90	Qwen3.8-27B	92.9%	estimated ± 0.8 pp, medium confidence
91	GPT-5.1	92.9%	estimated ± 1.0 pp, medium confidence
92	Seed 2.1 Pro	92.8%	estimated ± 0.7 pp, medium confidence
93	Ling 3.0 Flash VL	92.8%	estimated ± 1.0 pp, medium confidence
94	MiMo-V2-Omni	92.6%	estimated ± 1.0 pp, medium confidence
95	A.X K2	92.6%	estimated ± 0.8 pp, medium confidence
96	GPT-5.1-Codex	92.5%	estimated ± 1.0 pp, medium confidence
97	GPT-5.1-Codex-Max	92.5%	estimated ± 1.0 pp, medium confidence
98	GLM-5V-Turbo	92.5%	estimated ± 1.0 pp, medium confidence
99	Nemotron 3 Ultra	92.4%	estimated ± 0.8 pp, medium confidence
100	Gemma 4 31B	92.3%	estimated ± 0.8 pp, medium confidence
101	GPT-5 (high)	92.3%	estimated ± 1.0 pp, medium confidence
102	GPT-5 (medium)	92.3%	estimated ± 1.0 pp, medium confidence
103	Claude 4.1 Opus Thinking	92.3%	estimated ± 1.0 pp, medium confidence
104	MiniMax M2.7	92.2%	estimated ± 1.0 pp, medium confidence
105	Kimi K2.5	92.2%	estimated ± 0.7 pp, medium confidence
106	Claude Opus 4.5	92.2%	measured
107	Command A+	92.2%	estimated ± 1.0 pp, medium confidence
108	Grok 4	92.1%	estimated ± 1.0 pp, medium confidence
109	Ornith-1.5-35B-A3B	92.1%	estimated ± 0.8 pp, medium confidence
110	Hy3 Preview	92.1%	estimated ± 0.8 pp, medium confidence
111	Gemini 3.5 Flash-Lite	92.0%	estimated ± 1.0 pp, medium confidence
112	o3-pro	91.9%	estimated ± 1.0 pp, medium confidence
113	GLM-4.7	91.9%	estimated ± 0.8 pp, medium confidence
114	Kimi K2.5 (Reasoning)	91.8%	estimated ± 0.9 pp, medium confidence
115	Qwen3.5 397B (Reasoning)	91.8%	estimated ± 1.0 pp, medium confidence
116	Seed 2.1 Turbo	91.5%	estimated ± 0.7 pp, medium confidence
117	Qwen3.6-27B	91.4%	measured
118	Qwen3.5-122B-A10B	91.4%	estimated ± 0.7 pp, medium confidence
119	Grok 4.1 Fast (Reasoning)	91.3%	estimated ± 1.0 pp, medium confidence
120	o3	91.3%	estimated ± 1.0 pp, medium confidence
121	GLM-5	91.3%	estimated ± 0.7 pp, medium confidence
122	Solar Open 2	91.1%	estimated ± 1.0 pp, medium confidence
123	Step 3.7 Flash	91.0%	estimated ± 1.0 pp, medium confidence
124	Ling 3.0 Flash	90.9%	estimated ± 0.8 pp, medium confidence
125	Ternary Bonsai 2 27B	90.8%	estimated ± 0.9 pp, low confidence
126	Qwen3.5-27B	90.7%	estimated ± 0.7 pp, medium confidence
127	Claude 4.1 Opus	90.5%	estimated ± 1.0 pp, medium confidence
128	Grok 4 Fast (Reasoning)	90.2%	estimated ± 1.0 pp, low confidence
129	Gemini 3 Flash	90.2%	estimated ± 1.0 pp, low confidence
130	Qwen3.6-35B-A3B	90.0%	measured
131	Muse Glimmer 30B	90.0%	estimated ± 1.0 pp, low confidence
132	MAI-Thinking-1	89.8%	estimated ± 0.9 pp, low confidence
133	Qwen3.5-35B-A3B	89.8%	estimated ± 0.7 pp, low confidence
134	Ling 3.0 Flash FP8	89.7%	estimated ± 0.9 pp, low confidence
135	MiMo-V2-Flash	89.5%	estimated ± 0.9 pp, low confidence
136	Claude 4 Sonnet	89.5%	estimated ± 1.0 pp, low confidence
137	Qwen3 235B 2507	89.4%	estimated ± 0.7 pp, low confidence
138	Claude Sonnet 4.5	89.4%	estimated ± 0.9 pp, low confidence
139	DeepSeek V3.2	89.1%	estimated ± 1.0 pp, low confidence
140	Qwen3 Max	88.9%	estimated ± 1.0 pp, low confidence
141	Ornith-1.5-9B	88.7%	estimated ± 0.8 pp, low confidence
142	GLM-4.6	88.4%	estimated ± 1.0 pp, low confidence
143	K-Exaone	88.0%	estimated ± 1.0 pp, low confidence
144	Mistral Medium 3.5 128B	87.8%	estimated ± 1.0 pp, low confidence
145	Grok Code Fast 1	87.7%	estimated ± 1.0 pp, low confidence
146	DeepSeek V3.1	87.5%	estimated ± 1.0 pp, low confidence
147	DeepSeek V3.1 (Reasoning)	87.3%	estimated ± 1.0 pp, low confidence
148	DeepSeek-R1	87.0%	estimated ± 1.0 pp, low confidence
149	o1-pro	86.7%	estimated ± 0.9 pp, low confidence
150	Nemotron 3 Super 100B	86.7%	estimated ± 1.0 pp, low confidence
151	Kimi K2	86.6%	estimated ± 1.0 pp, low confidence
152	Gemma 4 12B	86.6%	estimated ± 0.9 pp, low confidence
153	Gemini 2.5 Pro	86.4%	estimated ± 0.8 pp, low confidence
154	Mercury 2.5	86.2%	estimated ± 1.0 pp, low confidence
155	LongCat-Flash-Lite-Sparse	85.8%	measured
156	o3-mini	85.6%	estimated ± 0.9 pp, low confidence
157	GPT-OSS 120B	85.4%	estimated ± 1.0 pp, low confidence
158	K-EXAONE 2.0	85.2%	estimated ± 0.8 pp, low confidence
159	o1-preview	85.2%	estimated ± 1.0 pp, low confidence
160	Grok 4.1 Fast	85.1%	estimated ± 1.0 pp, low confidence
161	Mistral Small 4	85.1%	estimated ± 1.0 pp, low confidence
162	Mistral Small 4 (Reasoning)	85.1%	estimated ± 1.0 pp, low confidence
163	GLM-4.5-Air	84.9%	estimated ± 1.0 pp, low confidence
164	Ling 3.0 Tiny	84.8%	estimated ± 1.0 pp, low confidence
165	o1	84.6%	estimated ± 0.9 pp, low confidence
166	Nemotron 3.5 Lightning 30B A3B NVFP4	84.5%	estimated ± 0.9 pp, low confidence
167	Trinity-Large-Preview	84.5%	estimated ± 1.0 pp, low confidence
168	Trinity-Large-Thinking	84.5%	estimated ± 1.0 pp, low confidence
169	Llama 4 Maverick	83.4%	estimated ± 1.0 pp, low confidence
170	North Mini Code	83.3%	estimated ± 1.0 pp, low confidence
171	Gemini 2.5 Flash	83.2%	estimated ± 1.0 pp, low confidence
172	DeepSeek V3 0324	83.0%	estimated ± 1.0 pp, low confidence
173	Mistral Large 3	82.3%	estimated ± 1.0 pp, low confidence
174	Nemotron 3 Nano Omni 30B A3B	82.3%	estimated ± 0.9 pp, low confidence
175	Gemma 4 26B A4B	82.1%	estimated ± 0.8 pp, low confidence
176	Mistral Medium 3	82.0%	estimated ± 1.0 pp, low confidence
177	GPT-OSS 20B	81.8%	estimated ± 1.0 pp, low confidence
178	Nemotron 3 Nano 30B	81.7%	estimated ± 1.0 pp, low confidence
179	Sarvam 105B	81.5%	estimated ± 1.0 pp, low confidence
180	ZAYA1-8B	81.5%	estimated ± 0.9 pp, low confidence
181	Claude 3 Opus	81.4%	estimated ± 1.0 pp, low confidence
182	GPT-4o	80.9%	estimated ± 1.0 pp, low confidence
183	LFM2.5-2.6B	80.8%	estimated ± 1.0 pp, low confidence
184	DeepSeek R1 Distill Qwen 32B	80.8%	estimated ± 1.0 pp, low confidence
185	Llama 4 Scout	80.2%	estimated ± 1.0 pp, low confidence
186	Gemini 1.5 Pro	79.8%	estimated ± 1.0 pp, low confidence
187	Solar Pro 3	79.6%	estimated ± 1.0 pp, low confidence
188	Qwen3-Omni-30B-A3B-Thinking	79.5%	estimated ± 1.0 pp, low confidence
189	Ultravox v0.6 Llama 3.3 70B	79.3%	estimated ± 1.0 pp, low confidence
190	Mistral Large 2	79.1%	estimated ± 1.0 pp, low confidence
191	Nemotron Ultra 253B	79.0%	estimated ± 1.0 pp, low confidence
192	Llama 3.1 405B	78.5%	estimated ± 1.0 pp, low confidence
193	LFM2.5-8B-A1B	78.3%	estimated ± 1.0 pp, low confidence
194	Granite 4.2 30B	78.3%	estimated ± 0.9 pp, low confidence
195	GPT-4.1	78.2%	estimated ± 0.9 pp, low confidence
196	GPT-4 Turbo	77.9%	estimated ± 1.0 pp, low confidence
197	Solar Pro 2	77.8%	estimated ± 1.0 pp, low confidence
198	Nova Pro	77.7%	estimated ± 1.0 pp, low confidence
199	Qwen2.5 Coder 32B Instruct	77.1%	estimated ± 1.0 pp, low confidence
200	GPT-4o mini	76.9%	estimated ± 1.0 pp, low confidence
201	GPT-4.1 mini	76.7%	estimated ± 0.9 pp, low confidence
202	Granite 4.2 8B	76.6%	estimated ± 0.9 pp, low confidence
203	Sarvam 30B	76.6%	estimated ± 1.0 pp, low confidence
204	MiniCPM5-2B	76.4%	estimated ± 0.7 pp, low confidence
205	Celeris-1	76.0%	estimated ± 1.0 pp, low confidence
206	Exaone 4.0 32B	75.9%	estimated ± 1.0 pp, low confidence
207	Qwen3-Omni-30B-A3B-Instruct	75.0%	estimated ± 1.0 pp, low confidence
208	Phi-4	74.7%	estimated ± 1.0 pp, low confidence
209	Phi-4 Multimodal Instruct	74.3%	estimated ± 1.0 pp, low confidence
210	Claude 3 Haiku	73.5%	estimated ± 1.0 pp, low confidence
211	Claude 3.5 Sonnet	73.0%	estimated ± 0.9 pp, low confidence
212	DeepSeek V3	72.8%	estimated ± 0.9 pp, low confidence
213	Ling 2.6 Flash	72.7%	estimated ± 0.9 pp, low confidence
214	Gemini 1.0 Pro	72.6%	estimated ± 1.0 pp, low confidence
215	Gemma 4 E4B	72.4%	estimated ± 0.9 pp, low confidence
216	Exaone 4.0 1.2B	72.2%	estimated ± 1.0 pp, low confidence
217	Granite-4.0-H-1B	72.1%	estimated ± 1.0 pp, low confidence
218	Mellum2-12B-A2.5B-Thinking	71.6%	estimated ± 0.9 pp, low confidence
219	ZAYA1-74B-Preview	71.4%	estimated ± 0.9 pp, low confidence
220	Gemma 3 27B	70.7%	estimated ± 1.0 pp, low confidence
221	Granite-4.0-350M	70.6%	estimated ± 1.0 pp, low confidence
222	Granite-4.0-H-350M	70.6%	estimated ± 1.0 pp, low confidence
223	LFM2.5-VL-1.6B-Extract	70.6%	estimated ± 1.0 pp, low confidence
224	Granite 4.2 3B	69.4%	estimated ± 0.9 pp, low confidence
225	GPT-4.1 nano	65.5%	estimated ± 0.9 pp, low confidence
226	Gemma 4 E2B	59.2%	estimated ± 0.9 pp, low confidence
227	Soofi S 30B-A3B	59.2%	estimated ± 0.9 pp, low confidence
228	MiniCPM5-1B	58.0%	estimated ± 0.7 pp, low confidence
229	Mellum2-12B-A2.5B-Instruct	56.8%	estimated ± 0.9 pp, low confidence
230	LFM2.5-VL-450M	39.8%	estimated ± 0.9 pp, low confidence
231	LFM2.5-230M	39.5%	estimated ± 0.9 pp, low confidence

C-Eval leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General