Vision & documents

BabyVision w/ Python leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap
License: https://creativecommons.org/licenses/by-nc/4.0/

As of 2026-11-07, the highest measured score on BabyVision w/ Python is 95.4% by Kimi K3.1 Pro. 77 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Opus 5.5	95.4%	estimated ± 4.0 pp, low confidence
2	Kimi K3.1 Pro	95.4%	measured
3	GPT-6 Astra	95.0%	estimated ± 4.0 pp, low confidence
4	GPT-6.1 Sol	94.4%	estimated ± 4.0 pp, low confidence
5	Claude Mythos 5	94.3%	estimated ± 3.3 pp, medium confidence
6	Inkling 2 Small	93.7%	measured
7	Claude Opus 5	93.7%	estimated ± 4.0 pp, low confidence
8	GPT-6.1 Terra	93.7%	estimated ± 4.0 pp, low confidence
9	Gemini 3.5 Flash	93.4%	estimated ± 4.0 pp, low confidence
10	GPT-5.6 Sol	92.9%	estimated ± 4.0 pp, low confidence
11	Gemini 3.6 Flash	92.8%	estimated ± 4.0 pp, low confidence
12	GPT-6 Sol	92.6%	estimated ± 4.0 pp, low confidence
13	Qwen3.8 Max Preview	92.5%	estimated ± 4.0 pp, low confidence
14	DeepSeek V4.1 Flash 1017	92.5%	measured
15	Gemini 3.1 Pro	92.3%	estimated ± 4.0 pp, low confidence
16	Muse Spark Open	91.4%	estimated ± 4.0 pp, low confidence
17	Qwen3.8 Max	91.3%	measured
18	GPT-5.6 Terra	91.3%	estimated ± 4.0 pp, low confidence
19	Muse Spark	91.1%	estimated ± 4.0 pp, low confidence
20	Qwen3.7 Plus	91.1%	estimated ± 4.0 pp, low confidence
21	Grok 4.5	91.1%	estimated ± 4.0 pp, low confidence
22	MiniMax M3.1 Flash	91.0%	measured
23	Gemini 3 Pro	91.0%	estimated ± 4.0 pp, low confidence
24	GPT-5.5	90.8%	estimated ± 4.0 pp, low confidence
25	GPT-6 Luna	90.7%	estimated ± 4.0 pp, low confidence
26	Kimi K2.6	90.5%	estimated ± 4.0 pp, low confidence
27	Apodex 1.1	90.3%	estimated ± 4.0 pp, low confidence
28	Apodex 1.1 Mini	90.3%	estimated ± 4.0 pp, low confidence
29	Gemini 3.8 Flash	90.3%	estimated ± 3.3 pp, medium confidence
30	Gemini 3.5 Flash-Lite	90.2%	estimated ± 4.0 pp, low confidence
31	Ling 3.0 Flash VL	90.2%	estimated ± 4.0 pp, low confidence
32	Gemini 3 Flash	90.0%	estimated ± 4.0 pp, low confidence
33	GPT-5.6 Luna	90.0%	estimated ± 4.0 pp, low confidence
34	MiniMax M3	90.0%	estimated ± 4.0 pp, low confidence
35	GPT-5.3 Codex	89.9%	estimated ± 4.0 pp, low confidence
36	GPT-5.4	89.8%	estimated ± 4.0 pp, low confidence
37	Grok 4.3	89.7%	estimated ± 4.0 pp, low confidence
38	Qwen3.6 Plus	89.6%	estimated ± 4.0 pp, low confidence
39	Claude Sonnet 5	89.1%	estimated ± 4.0 pp, low confidence
40	Qwen3.5 397B	89.1%	estimated ± 4.0 pp, low confidence
41	Qwen3.5 397B (Reasoning)	89.1%	estimated ± 4.0 pp, low confidence
42	Claude Opus 4.7	88.6%	estimated ± 4.0 pp, low confidence
43	Step 5 Preview	88.6%	estimated ± 4.0 pp, low confidence
44	GPT-5.2-Codex	88.5%	estimated ± 4.0 pp, low confidence
45	GPT-5.1	88.0%	estimated ± 4.0 pp, low confidence
46	Claude Opus 4.6 (Adaptive)	87.9%	estimated ± 4.0 pp, low confidence
47	Kimi K2.5	87.9%	estimated ± 4.0 pp, low confidence
48	Kimi K2.5 (Reasoning)	87.9%	estimated ± 4.0 pp, low confidence
49	Step 3.7 Flash	87.8%	estimated ± 4.0 pp, low confidence
50	Qwen3.8-Flash-Next	87.8%	estimated ± 3.3 pp, medium confidence
51	Qwen3.5-122B-A10B	87.7%	estimated ± 4.0 pp, low confidence
52	Qwen3.5-27B	87.7%	estimated ± 4.0 pp, low confidence
53	Qwen3.6-35B-A3B	87.7%	estimated ± 4.0 pp, low confidence
54	Gemini 3.7 Flash	87.6%	estimated ± 3.3 pp, medium confidence
55	Gemini 2.5 Pro	87.6%	estimated ± 4.0 pp, low confidence
56	Qwen3.6-27B	87.4%	estimated ± 4.0 pp, low confidence
57	GPT-5 (medium)	87.2%	estimated ± 4.0 pp, low confidence
58	Muse Glimmer 30B	87.2%	estimated ± 4.0 pp, low confidence
59	GPT-5 (high)	87.1%	estimated ± 4.0 pp, low confidence
60	Claude Opus 4.5 Thinking	87.0%	estimated ± 4.0 pp, low confidence
61	Inkling-Small	87.0%	estimated ± 4.0 pp, low confidence
62	Inkling	86.7%	estimated ± 4.0 pp, low confidence
63	Gemma 4 31B	86.6%	estimated ± 4.0 pp, low confidence
64	GPT-5.4 mini	86.5%	estimated ± 4.0 pp, low confidence
65	MiMo-V2.6-Flash	86.4%	estimated ± 4.0 pp, low confidence
66	GLM-5V-Turbo	86.2%	estimated ± 4.0 pp, low confidence
67	Qwen3.5-35B-A3B	86.1%	estimated ± 4.0 pp, low confidence
68	Claude Opus 4.6	86.0%	estimated ± 4.0 pp, low confidence
69	GPT-5.1-Codex	86.0%	estimated ± 4.0 pp, low confidence
70	GPT-5.1-Codex-Max	86.0%	estimated ± 4.0 pp, low confidence
71	Qwen3.8-Omni-Flash	85.9%	estimated ± 3.3 pp, low confidence
72	Kimi K3	85.7%	measured
73	Qwen3.8-27B	85.6%	measured
74	dots3-note Preview	85.2%	estimated ± 3.3 pp, low confidence
75	Claude Opus 4.5	85.1%	estimated ± 4.0 pp, low confidence
76	Claude Sonnet 4.6	84.7%	estimated ± 4.0 pp, low confidence
77	o3	84.3%	estimated ± 4.0 pp, low confidence
78	MiMo-V2-Omni	84.2%	estimated ± 4.0 pp, low confidence
79	Gemma 4 12B	84.1%	estimated ± 4.0 pp, low confidence
80	Gemma 4 26B A4B	83.7%	estimated ± 4.0 pp, low confidence
81	o4-mini	83.7%	estimated ± 4.0 pp, low confidence
82	Grok 4	83.4%	estimated ± 4.0 pp, low confidence
83	Claude Opus 4.7 (Adaptive)	83.3%	estimated ± 3.3 pp, low confidence
84	Claude Opus 4.8	80.3%	estimated ± 3.3 pp, low confidence

BabyVision w/ Python leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General