BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases12

LiveBench 1.0

Average of LiveBench 1.0's seven category scores, frozen at its final release (25 June 2026). LiveBench 2.0 replaced it.

As of 24 Oct 2026, Claude Fable 5.1 leads LiveBench 1.0 on BenchLeader with 83.4%, ahead of Claude Opus 5.5 at 83.2%, across 67 model configurations with a published result.

Published by
LiveBench
Category
Composite
Index weight
Reference only
Models
67
Data as of
24 Oct 2026
Newest result
25 Jun 2026

The publisher has posted no new results on this board since 25 Jun 2026. BenchLeader re-reads it every day, so newer models appear here as soon as they are published.

LiveBench (livebench.ai), White et al. 2024.

This board is frozen. A different, newer test now carries its name, and its scores are on LiveBench 2.0.

What the test looks like

LiveBench 1.0 was a contamination-focused benchmark refreshed by release: freshly written questions across reasoning, coding, agentic coding, mathematics, data analysis, language and instruction following.

How it is scored

The average of LiveBench 1.0's seven category scores, from its final release (25 June 2026). LiveBench 2.0 replaced it with a different set of suites, so these scores no longer update.

What to keep in mind

Not comparable with the LiveBench 2.0 overall score, which averages different suites. Models released after June 2026 are not on it.

Top 15
  1. 1Claude Fable 5.1 (max)83.4%
  2. 2Claude Opus 5.5 (max)83.2%
  3. 3Claude Fable 5 (max)83.0%
  4. 4GPT-6 Astra (max)82.2%
  5. 5Claude Opus 5.5 (xhigh)82.1%
  6. 6GPT-6.1 Sol (max)81.6%
  7. 7Muse Spark 1.3 (xhigh)81.6%
  8. 8Muse Wonder (max)81.2%
  9. 9GPT-6.1 Sol (xhigh)81.1%
  10. 10DeepSeek V4.1 Flash (max)81.1%
  11. 11GPT-5.6 Sol (max)81.0%
  12. 12GPT-5.5 (xhigh)80.2%
  13. 13Claude Opus 5 (max)80.1%
  14. 14GPT-6 Sol (max)79.3%
  15. 15Kimi K379.2%
67 of 67
#
1Claude Fable 5.1max83.4%70.52026-06-25
2Claude Opus 5.5max83.2%72.62026-06-25
3Claude Fable 5max83.0%69.32026-06-25
4GPT-6 Astramax82.2%70.42026-06-25
5Claude Opus 5.5xhigh82.1%70.02026-06-25
6GPT-6.1 Solmax81.6%69.92026-06-25
7Muse Spark 1.3xhigh81.6%64.92026-06-25
8Muse Wondermax81.2%67.52026-06-25
9GPT-6.1 Solxhigh81.1%67.72026-06-25
10DeepSeek V4.1 Flashmax81.1%61.22026-06-25
11GPT-5.6 Solmax81.0%68.02026-06-25
12GPT-5.5xhigh80.2%66.12026-06-25
13Claude Opus 5max80.1%68.82026-06-25
14GPT-6 Solmax79.3%66.42026-06-25
15Kimi K379.2%59.12026-06-25
16Gemini 3.7 Flashhigh78.8%63.52026-06-25
17Qwen3.8 Maxmax78.5%63.32026-06-25
18Grok 4.678.0%57.82026-06-25
19GPT-5.4xhigh78.0%64.72026-06-25
20Muse Spark 1.2xhigh78.0%63.22026-06-25
21GPT-5.6 Terramax77.9%64.42026-06-25
22Claude Sonnet 5.5xhigh77.8%67.92026-06-25
23DeepSeek V4 Pro77.4%53.52026-06-25
24Smaug Flash77.4%–2026-06-25
25Grok 4.7xhigh77.4%61.62026-06-25
26Gemini 3.1 Prohigh77.0%60.32026-06-25
27Smaug Mini76.9%–2026-06-25
28Deepseek v4 Flash Vision76.8%–2026-06-25
29Claude Opus 4.7xhigh76.5%57.42026-06-25
30Claude Opus 4.8max76.2%63.32026-06-25
31Qwen3.8-Flash-Next76.2%60.22026-06-25
32GLM 5.376.1%–2026-06-25
33Claude Sonnet 5xhigh76.0%59.02026-06-25
34Gemini 3.8 Flashhigh75.8%64.42026-06-25
35Grok 4.575.8%59.02026-06-25
36Claude Sonnet 5.5max75.7%69.32026-06-25
37Muse Spark 1.1xhigh75.3%61.62026-06-25
38Qwen3.8 27B75.3%54.62026-06-25
39Gemini 3.5 Flashhigh74.6%63.12026-06-25
40GPT-5.2high74.6%56.02026-06-25
41Claude Opus 4.6high74.5%62.72026-06-25
42DeepSeek V4 Flash74.2%52.22026-06-25
43GPT-5.2-Codex74.0%57.32026-06-25
44Gemini 3.6 Flashhigh73.6%62.02026-06-25
45GPT-5.6 Lunamax73.6%59.92026-06-25
46GLM 5.273.2%52.02026-06-25
47Qwen3.7max73.1%60.62026-06-25
48Claude Sonnet 4.6medium73.0%–2026-06-25
49Claude Opus 4.5high72.6%59.22026-06-25
50Claude Haiku 5.5xhigh72.1%61.52026-06-25
51GPT-6 Lunamax72.0%60.02026-06-25
52Mistral Large 4high71.8%49.82026-06-25
53GLM 5.3 Flash71.6%62.62026-06-25
54Kimi K2.6thinking70.5%59.42026-06-25
55GPT-5.4 Nanoxhigh69.6%54.22026-06-25
56Ox Alphamax69.2%–2026-06-25
57Qwen3.6 Plus68.9%58.82026-06-25
58Kimi K2.7 Code68.4%55.82026-06-25
59Claude Haiku 5.5max67.9%60.52026-06-25
60Grok Build 0.167.8%54.82026-06-25

More composite benchmarks · All LLM benchmarks

Cite as: BenchLeader, “LiveBench 1.0 leaderboard”, data as of 24 Oct 2026, https://www.benchleader.com/benchmarks/livebench_global.

LiveBench 1.0: questions

What does LiveBench 1.0 measure?
LiveBench 1.0 was a contamination-focused benchmark refreshed by release: freshly written questions across reasoning, coding, agentic coding, mathematics, data analysis, language and instruction following. Scores are reported in percent of tasks solved; higher is better.
Which AI model leads LiveBench 1.0?
Claude Fable 5.1 leads LiveBench 1.0 with 83.4% as of 24 Oct 2026, ahead of Claude Opus 5.5 at 83.2%.
How many models have LiveBench 1.0 results?
67 model configurations have a LiveBench 1.0 result on BenchLeader, all taken from LiveBench.
Who runs LiveBench 1.0 and how often is it updated?
LiveBench 1.0 is published by LiveBench. BenchLeader re-reads the published results once a day and records the date each result was published.
Does LiveBench 1.0 count toward the BenchLeader Index?
No. LiveBench 1.0 is shown for reference but left out of the composite index.