livebench math integrals with game 2026_06_25 Benchmark: Scores and Sources

Published result; benchmark version and evaluation conditions remain in the id and result note.

pillar: math · weight 1 within pillar · unit: % (higher is better) · official board ↗

Row	Model	Result	Normalized (0–100)	Confidence
1	Claude Sonnet 5.5 Anthropic	100%	100.0	93% confidence 93 percent, High
2	GPT-5.6 Sol OpenAI	100%	100.0	100% confidence 100 percent, Full
3	GPT-6 Astra OpenAI	100%	100.0	100% confidence 100 percent, Full
4	Claude Fable 5.1 Anthropic	99%	99.0	100% confidence 100 percent, Full
5	Claude Opus 5.5 Anthropic	99%	99.0	100% confidence 100 percent, Full
6	GPT-5.5 OpenAI	99%	99.0	100% confidence 100 percent, Full
7	GPT-6 Sol OpenAI	99%	99.0	100% confidence 100 percent, Full
8	GPT-6.1 Sol OpenAI	99%	99.0	100% confidence 100 percent, Full
9	Claude Fable 5 Anthropic	97%	97.0	100% confidence 100 percent, Full
10	Claude Opus 5 Anthropic	97%	97.0	100% confidence 100 percent, Full
11	Muse Spark 1.3 Meta	96%	96.0	93% confidence 93 percent, High
12	Grok 4.7 xAI	96%	96.0	100% confidence 100 percent, Full
13	GPT-5.6 Terra OpenAI	95%	95.0	100% confidence 100 percent, Full
14	DeepSeek V4 Pro 0813 DeepSeek	94%	94.0	69% confidence 69 percent, Medium
15	GPT-5.4 OpenAI	93%	93.0	100% confidence 100 percent, Full
16	GPT-5.2 OpenAI	92%	92.0	100% confidence 100 percent, Full
17	Muse Spark 1.2 Meta	91%	91.0	53% confidence 53 percent, Medium
18	Claude Sonnet 4.6 Anthropic	90%	90.0	100% confidence 100 percent, Full
19	DeepSeek V4.1 Flash DeepSeek	90%	90.0	69% confidence 69 percent, Medium
20	Mistral Large 4 mistral	90%	90.0	32% confidence 32 percent, Low
21	Claude Opus 4.8 Anthropic	89%	89.0	100% confidence 100 percent, Full
22	Claude Sonnet 5 Anthropic	88%	88.0	93% confidence 93 percent, High
23	Gemini 3.7 Flash Google	88%	88.0	100% confidence 100 percent, Full
24	Grok 4.6 xAI	85%	85.0	100% confidence 100 percent, Full
25	Claude Opus 4.7 Anthropic	84%	84.0	100% confidence 100 percent, Full
26	Muse Spark 1.1 Meta	83%	83.0	61% confidence 61 percent, Medium
27	GPT-5.4 nano OpenAI	82.8%	82.8	100% confidence 100 percent, Full
28	Qwen3.8 Max Alibaba / Qwen	81%	81.0	85% confidence 85 percent, High
29	Gemini 3.8 Flash Google	80%	80.0	100% confidence 100 percent, Full
30	Nemotron 3 Ultra 550B A55B nvidia	80%	80.0	100% confidence 100 percent, Full
31	Claude Opus 4.5 Anthropic	78%	78.0	100% confidence 100 percent, Full
32	DeepSeek V4 Pro DeepSeek	78%	78.0	85% confidence 85 percent, High
33	Gemini 3.1 Pro Preview Google	78%	78.0	100% confidence 100 percent, Full
34	Grok 4.5 xAI	78%	78.0	100% confidence 100 percent, Full
35	DeepSeek V4 Flash Vision Exp DeepSeek	76%	76.0	21% confidence 21 percent, Low
36	GLM-5.2 Z.ai	76%	76.0	100% confidence 100 percent, Full
37	GPT-5.2 Codex OpenAI	75%	75.0	43% confidence 43 percent, Low
38	GPT-6 Luna OpenAI	74%	74.0	100% confidence 100 percent, Full
39	Claude Opus 4.6 Anthropic	73%	73.0	100% confidence 100 percent, Full
40	Inkling thinkingmachines	72%	72.0	100% confidence 100 percent, Full
41	GPT-5.6 Luna OpenAI	70%	70.0	100% confidence 100 percent, Full
42	GLM-5.3 Z.ai	69%	69.0	93% confidence 93 percent, High
43	Gemini 3.5 Flash Google	68%	68.0	100% confidence 100 percent, Full
44	Claude Haiku 5.5 Anthropic	67%	67.0	100% confidence 100 percent, Full
45	Grok Build 0.1 xAI	66%	66.0	16% confidence 16 percent, Low
46	Qwen3.8 27B Alibaba / Qwen	65%	65.0	69% confidence 69 percent, Medium
47	DeepSeek V4 Flash 0731 DeepSeek	65%	65.0	69% confidence 69 percent, Medium
48	Gemini 3.6 Flash Google	63%	63.0	100% confidence 100 percent, Full
49	Grok 4.3 xAI	63%	63.0	80% confidence 80 percent, High
50	Qwen3.6 Plus Alibaba / Qwen	62%	62.0	80% confidence 80 percent, High
51	Qwen3.7 Max Alibaba / Qwen	59%	59.0	53% confidence 53 percent, Medium
52	Qwen3.8 Flash Next Alibaba / Qwen	59%	59.0	21% confidence 21 percent, Low
53	MiniMax-M3 minimax	56%	56.0	100% confidence 100 percent, Full
54	Kimi K2.6 Moonshot AI	54%	54.0	85% confidence 85 percent, High
55	Kimi K3 Moonshot AI	54%	54.0	100% confidence 100 percent, Full
56	Qwen3.6 27B Alibaba / Qwen	52%	52.0	53% confidence 53 percent, Medium
57	GLM-5.3-Flash Z.ai	45%	45.0	100% confidence 100 percent, Full
58	DeepSeek V4 Flash DeepSeek	37%	37.0	53% confidence 53 percent, Medium
59	Kimi K2.7 Code Moonshot AI	37%	37.0	48% confidence 48 percent, Low
60	GPT-5.4 mini OpenAI	37%	37.0	100% confidence 100 percent, Full
61	Gemini 3.5 Flash Lite Google	26%	26.0	100% confidence 100 percent, Full

Results are as published by the source behind each value (hover or tap the number). Benchmark scores are shown individually for every source/variant (a model may have multiple rows), and vary by version, harness and date; the normalized column uses the method’s fixed 0–100 scales and feeds the math pillar of the SI Score.