---
title: "LLM Leaderboard Top 10: #1 Leads #10 by 18 Points"
date: 2026-09-14
url: https://modevc.com/charts/llm-leaderboard-top-10-gap/
keywords: ["llm leaderboard top 10", "llm leaderboard", "llm benchmark"]
publisher: Mode VC
---

# LLM Leaderboard Top 10: #1 Leads #10 by 18 Points

The #10 model on the Arena board would now win 47% of blind votes against #1. When the top ten are this close, quality stops deciding the sale.

## Chart: The gap between Arena's #1 and #10 models fell from 148 points in Feb 2024 to 18. #10 now wins 47% of head-to-head votes against #1, up from 30%.

Rating of the #1 model minus the #10 and #20 models, Arena text board (overall), Arena points, 9 Jan 2024 – 13 Sept 2026.

| Series | Date | Value | Note |
|---|---|---|---|
| #1 minus #10 | 2024-01-09 | 97 | vicuna-33b |
| #1 minus #10 | 2024-01-18 | 97 | vicuna-33b |
| #1 minus #10 | 2024-01-25 | 95 | vicuna-33b |
| #1 minus #10 | 2024-02-02 | 148 | wizardlm-70b |
| #1 minus #10 | 2024-02-15 | 142 | yi-34b-chat |
| #1 minus #10 | 2024-03-05 | 139 | gemini-pro |
| #1 minus #10 | 2024-03-07 | 133 | gemini-pro-dev-api |
| #1 minus #10 | 2024-03-13 | 133 | gemini-pro-dev-api |
| #1 minus #10 | 2024-03-26 | 108 | mistral-medium |
| #1 minus #10 | 2024-03-27 | 106 | command-r |
| #1 minus #10 | 2024-04-03 | 110 | mistral-medium |
| #1 minus #10 | 2024-04-09 | 103 | qwen1.5-72b-chat |
| #1 minus #10 | 2024-04-10 | 103 | qwen1.5-72b-chat |
| #1 minus #10 | 2024-04-11 | 102 | mistral-large-2402 |
| #1 minus #10 | 2024-04-13 | 101 | mistral-large-2402 |
| #1 minus #10 | 2024-04-18 | 94 | gpt-4-0613 |
| #1 minus #10 | 2024-04-19 | 94 | gpt-4-0613 |
| #1 minus #10 | 2024-04-22 | 94 | gpt-4-0613 |
| #1 minus #10 | 2024-04-26 | 93 | gpt-4-0613 |
| #1 minus #10 | 2024-05-01 | 94 | gpt-4-0613 |
| #1 minus #10 | 2024-05-08 | 86 | qwen1.5-110b-chat |
| #1 minus #10 | 2024-05-15 | 108 | claude-3-haiku-20240307 |
| #1 minus #10 | 2024-05-16 | 106 | claude-3-haiku-20240307 |
| #1 minus #10 | 2024-05-19 | 105 | claude-3-haiku-20240307 |
| #1 minus #10 | 2024-05-20 | 105 | claude-3-haiku-20240307 |
| #1 minus #10 | 2024-05-27 | 83 | claude-3-sonnet-20240229 |
| #1 minus #10 | 2024-06-02 | 83 | claude-3-sonnet-20240229 |
| #1 minus #10 | 2024-06-06 | 83 | claude-3-sonnet-20240229 |
| #1 minus #10 | 2024-06-11 | 83 | claude-3-sonnet-20240229 |
| #1 minus #10 | 2024-06-17 | 81 | llama-3-70b-instruct |
| #1 minus #10 | 2024-06-21 | 77 | nemotron-4-340b-instruct |
| #1 minus #10 | 2024-06-23 | 77 | glm-4-0520 |
| #1 minus #10 | 2024-06-26 | 79 | glm-4-0520 |
| #1 minus #10 | 2024-06-29 | 73 | gemma-2-27b-it |
| #1 minus #10 | 2024-07-06 | 70 | gemma-2-27b-it |
| #1 minus #10 | 2024-07-08 | 71 | gemma-2-27b-it |
| #1 minus #10 | 2024-07-16 | 71 | gemma-2-27b-it |
| #1 minus #10 | 2024-07-22 | 59 | gemini-1.5-flash-api-0514 |
| #1 minus #10 | 2024-07-25 | 59 | gemini-1.5-flash-api-0514 |
| #1 minus #10 | 2024-07-30 | 41 | athene-70b-0725 |
| #1 minus #10 | 2024-07-31 | 41 | gpt-4-0125-preview |
| #1 minus #10 | 2024-08-01 | 40 | gpt-4-0125-preview |
| #1 minus #10 | 2024-08-05 | 38 | claude-3-opus-20240229 |
| #1 minus #10 | 2024-08-06 | 37 | mistral-large-2407 |
| #1 minus #10 | 2024-08-13 | 38 | claude-3-opus-20240229 |
| #1 minus #10 | 2024-08-22 | 66 | gpt-4-1106-preview |
| #1 minus #10 | 2024-08-23 | 57 | gemini-1.5-pro-api-0514 |
| #1 minus #10 | 2024-08-27 | 57 | gemini-1.5-pro-api-0514 |
| #1 minus #10 | 2024-08-28 | 57 | gemini-1.5-pro-api-0514 |
| #1 minus #10 | 2024-09-04 | 56 | gemini-1.5-pro-api-0514 |
| #1 minus #10 | 2024-09-15 | 35 | gemini-1.5-pro-api-0514 |
| #1 minus #10 | 2024-09-17 | 91 | llama-3.1-405b-instruct-bf16 |
| #1 minus #10 | 2024-09-27 | 73 | llama-3.1-405b-instruct-fp8 |
| #1 minus #10 | 2024-10-07 | 69 | gemini-advanced-0514 |
| #1 minus #10 | 2024-10-15 | 66 | claude-3-5-sonnet-20240620 |
| #1 minus #10 | 2024-10-23 | 68 | claude-3-5-sonnet-20240620 |
| #1 minus #10 | 2024-10-28 | 63 | gemini-1.5-flash-002 |
| #1 minus #10 | 2024-11-04 | 62 | gemini-1.5-flash-002 |
| #1 minus #10 | 2024-11-12 | 62 | gemini-1.5-flash-002 |
| #1 minus #10 | 2024-11-13 | 61 | gemini-1.5-flash-002 |
| #1 minus #10 | 2024-11-20 | 63 | gemini-1.5-flash-002 |
| #1 minus #10 | 2024-11-21 | 62 | gemini-1.5-flash-002 |
| #1 minus #10 | 2024-11-22 | 61 | gpt-4o-mini-2024-07-18 |
| #1 minus #10 | 2024-12-01 | 61 | gpt-4o-mini-2024-07-18 |
| #1 minus #10 | 2024-12-05 | 62 | gpt-4o-mini-2024-07-18 |
| #1 minus #10 | 2024-12-10 | 61 | gpt-4o-mini-2024-07-18 |
| #1 minus #10 | 2024-12-15 | 61 | gpt-4o-mini-2024-07-18 |
| #1 minus #10 | 2024-12-18 | 61 | gpt-4o-mini-2024-07-18 |
| #1 minus #10 | 2024-12-22 | 58 | athene-v2-chat |
| #1 minus #10 | 2024-12-30 | 80 | qwen2.5-plus-1127 |
| #1 minus #10 | 2025-01-05 | 69 | qwen2.5-plus-1127 |
| #1 minus #10 | 2025-01-15 | 67 | claude-3-5-sonnet-20241022 |
| #1 minus #10 | 2025-01-19 | 68 | claude-3-5-sonnet-20241022 |
| #1 minus #10 | 2025-01-21 | 68 | claude-3-5-sonnet-20241022 |
| #1 minus #10 | 2025-01-22 | 68 | claude-3-5-sonnet-20241022 |
| #1 minus #10 | 2025-01-24 | 72 | gpt-4o-2024-05-13 |
| #1 minus #10 | 2025-01-28 | 73 | gpt-4o-2024-05-13 |
| #1 minus #10 | 2025-02-03 | 73 | grok-2-2024-08-13 |
| #1 minus #10 | 2025-02-05 | 60 | gemini-1.5-pro-002 |
| #1 minus #10 | 2025-02-06 | 58 | glm-4-plus-0111 |
| #1 minus #10 | 2025-02-09 | 57 | step-2-16k-exp-202412 |
| #1 minus #10 | 2025-02-11 | 57 | step-2-16k-exp-202412 |
| #1 minus #10 | 2025-02-14 | 56 | glm-4-plus-0111 |
| #1 minus #10 | 2025-02-17 | 56 | o3-mini |
| #1 minus #10 | 2025-02-21 | 53 | gemini-2.0-flash-lite-preview-02-05 |
| #1 minus #10 | 2025-02-27 | 53 | qwen-plus-0125 |
| #1 minus #10 | 2025-03-03 | 101 | glm-4-plus-0111 |
| #1 minus #10 | 2025-03-11 | 88 | deepseek-v3 |
| #1 minus #10 | 2025-03-15 | 88 | deepseek-v3 |
| #1 minus #10 | 2025-03-17 | 88 | deepseek-v3 |
| #1 minus #10 | 2025-03-20 | 86 | deepseek-v3 |
| #1 minus #10 | 2025-03-25 | 85 | deepseek-v3 |
| #1 minus #10 | 2025-03-26 | 82 | o3-mini-high |
| #1 minus #10 | 2025-03-30 | 71 | o1-preview |
| #1 minus #10 | 2025-04-05 | 75 | o1-preview |
| #1 minus #10 | 2025-04-08 | 75 | o1-preview |
| #1 minus #10 | 2025-04-09 | 71 | o1-preview |
| #1 minus #10 | 2025-04-16 | 72 | o1-preview |
| #1 minus #10 | 2025-04-23 | 69 | o1-2024-12-17 |
| #1 minus #10 | 2025-05-05 | 62 | o1-2024-12-17 |
| #1 minus #10 | 2025-05-06 | 61 | o1-2024-12-17 |
| #1 minus #10 | 2025-05-11 | 62 | o4-mini-2025-04-16 |
| #1 minus #10 | 2025-05-19 | 63 | o1-2024-12-17 |
| #1 minus #10 | 2025-05-20 | 63 | o1-2024-12-17 |
| #1 minus #10 | 2025-05-22 | 63 | o1-2024-12-17 |
| #1 minus #10 | 2025-06-01 | 63 | o1-2024-12-17 |
| #1 minus #10 | 2025-06-03 | 63 | o1-2024-12-17 |
| #1 minus #10 | 2025-06-05 | 62 | o1-2024-12-17 |
| #1 minus #10 | 2025-06-11 | 62 | o1-2024-12-17 |
| #1 minus #10 | 2025-06-17 | 53 | claude-opus-4-20250514 |
| #1 minus #10 | 2025-06-18 | 53 | hunyuan-turbos-20250416 |
| #1 minus #10 | 2025-06-24 | 92 | deepseek-v3-0324 |
| #1 minus #10 | 2025-07-01 | 89 | deepseek-v3-0324 |
| #1 minus #10 | 2025-07-07 | 95 | deepseek-v3-0324 |
| #1 minus #10 | 2025-07-15 | 91 | gpt-4.1-2025-04-14 |
| #1 minus #10 | 2025-07-17 | 90 | gpt-4.1-2025-04-14 |
| #1 minus #10 | 2025-07-25 | 91 | gpt-4.1-2025-04-14 |
| #1 minus #10 | 2025-07-28 | 94 | deepseek-v3-0324 |
| #1 minus #10 | 2025-08-01 | 57 | qwen3-235b-a22b-thinking-2507 |
| #1 minus #10 | 2025-08-04 | 57 | qwen3-235b-a22b-thinking-2507 |
| #1 minus #10 | 2025-08-07 | 48 | grok-3-preview-02-24 |
| #1 minus #10 | 2025-08-14 | 46 | grok-3-preview-02-24 |
| #1 minus #10 | 2025-08-18 | 49 | claude-opus-4-1-20250805-thinking-16k |
| #1 minus #10 | 2025-08-21 | 47 | claude-opus-4-1-20250805-thinking-16k |
| #1 minus #10 | 2025-08-27 | 47 | deepseek-v3.1 |
| #1 minus #10 | 2025-08-28 | 47 | deepseek-v3.1 |
| #1 minus #10 | 2025-09-08 | 45 | o3-2025-04-16 |
| #1 minus #10 | 2025-09-17 | 49 | o3-2025-04-16 |
| #1 minus #10 | 2025-09-18 | 48 | claude-opus-4-1-20250805-thinking-16k |
| #1 minus #10 | 2025-09-30 | 44 | grok-4-0709 |
| #1 minus #10 | 2025-10-01 | 44 | grok-4-0709 |
| #1 minus #10 | 2025-10-03 | 41 | grok-3-preview-02-24 |
| #1 minus #10 | 2025-10-07 | 41 | grok-3-preview-02-24 |
| #1 minus #10 | 2025-10-16 | 42 | claude-sonnet-4-5-20250929-thinking-32k |
| #1 minus #10 | 2025-10-17 | 42 | claude-sonnet-4-5-20250929-thinking-32k |
| #1 minus #10 | 2025-11-05 | 43 | longcat-flash-chat |
| #1 minus #10 | 2025-11-06 | 41 | grok-3-preview-02-24 |
| #1 minus #10 | 2025-11-09 | 42 | grok-3-preview-02-24 |
| #1 minus #10 | 2025-11-16 | 67 | glm-4.5 |
| #1 minus #10 | 2025-11-17 | 69 | glm-4.5 |
| #1 minus #10 | 2025-11-18 | 64 | glm-4.5 |
| #1 minus #10 | 2025-11-20 | 62 | glm-4.5 |
| #1 minus #10 | 2025-11-26 | 57 | mistral-medium-2508 |
| #1 minus #10 | 2025-12-01 | 53 | mistral-large-3 |
| #1 minus #10 | 2025-12-04 | 49 | mistral-large-3 |
| #1 minus #10 | 2025-12-05 | 48 | qwen3-max-preview |
| #1 minus #10 | 2025-12-08 | 52 | mistral-large-3 |
| #1 minus #10 | 2025-12-11 | 54 | mistral-large-3 |
| #1 minus #10 | 2025-12-15 | 54 | mistral-large-3 |
| #1 minus #10 | 2025-12-16 | 48 | qwen3-max-preview |
| #1 minus #10 | 2025-12-18 | 48 | qwen3-max-preview |
| #1 minus #10 | 2025-12-19 | 47 | qwen3-max-preview |
| #1 minus #10 | 2025-12-21 | 44 | glm-4.6 |
| #1 minus #10 | 2025-12-23 | 44 | glm-4.6 |
| #1 minus #10 | 2025-12-30 | 45 | glm-4.6 |
| #1 minus #10 | 2026-01-09 | 44 | glm-4.6 |
| #1 minus #10 | 2026-01-27 | 39 | ernie-5.0-preview-1203 |
| #1 minus #10 | 2026-01-29 | 38 | grok-4.1 |
| #1 minus #10 | 2026-02-06 | 56 | claude-opus-4-5-20251101 |
| #1 minus #10 | 2026-02-09 | 58 | gpt-5.1-high |
| #1 minus #10 | 2026-02-10 | 56 | grok-4.1-thinking |
| #1 minus #10 | 2026-02-11 | 56 | grok-4.1-thinking |
| #1 minus #10 | 2026-02-16 | 56 | gemini-3-flash (thinking-minimal) |
| #1 minus #10 | 2026-02-17 | 56 | gemini-3-flash (thinking-minimal) |
| #1 minus #10 | 2026-02-19 | 55 | gemini-3-flash (thinking-minimal) |
| #1 minus #10 | 2026-02-20 | 54 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-02-21 | 55 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-02-22 | 55 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-02-24 | 54 | gemini-3-flash (thinking-minimal) |
| #1 minus #10 | 2026-02-25 | 54 | gemini-3-flash (thinking-minimal) |
| #1 minus #10 | 2026-02-26 | 50 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-03-03 | 54 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-03-04 | 55 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-03-05 | 53 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-03-06 | 54 | qwen3.5-397b-a17b |
| #1 minus #10 | 2026-03-11 | 52 | amazon-nova-experimental-chat-26-02-10 |
| #1 minus #10 | 2026-03-16 | 40 | dola-seed-2.0-preview |
| #1 minus #10 | 2026-03-18 | 42 | grok-4.20-beta-0309-reasoning |
| #1 minus #10 | 2026-03-19 | 42 | grok-4.20-beta-0309-reasoning |
| #1 minus #10 | 2026-03-20 | 43 | dola-seed-2.0-preview |
| #1 minus #10 | 2026-03-26 | 43 | grok-4.20-beta-0309-reasoning |
| #1 minus #10 | 2026-03-31 | 43 | dola-seed-2.0-preview |
| #1 minus #10 | 2026-04-02 | 42 | dola-seed-2.0-pro |
| #1 minus #10 | 2026-04-07 | 39 | gemini-2.5-pro |
| #1 minus #10 | 2026-04-09 | 40 | gemini-2.5-pro |
| #1 minus #10 | 2026-04-10 | 40 | gemini-2.5-pro |
| #1 minus #10 | 2026-04-14 | 39 | gemini-2.5-pro |
| #1 minus #10 | 2026-04-17 | 31 | glm-5.1 |
| #1 minus #10 | 2026-04-20 | 32 | gemini-3-flash |
| #1 minus #10 | 2026-04-22 | 33 | gemini-3-flash |
| #1 minus #10 | 2026-04-23 | 33 | gemini-3-flash |
| #1 minus #10 | 2026-04-27 | 29 | gpt-5.4-high |
| #1 minus #10 | 2026-04-28 | 29 | gpt-5.4-high |
| #1 minus #10 | 2026-04-29 | 31 | gpt-5.4-high |
| #1 minus #10 | 2026-04-30 | 31 | qwen3.5-max-preview |
| #1 minus #10 | 2026-05-01 | 31 | qwen3.5-max-preview |
| #1 minus #10 | 2026-05-06 | 30 | gpt-5.5-high |
| #1 minus #10 | 2026-05-07 | 31 | gpt-5.4-high |
| #1 minus #10 | 2026-05-12 | 30 | gpt-5.4-high |
| #1 minus #10 | 2026-05-14 | 29 | gpt-5.4-high |
| #1 minus #10 | 2026-05-18 | 29 | gpt-5.4-high |
| #1 minus #10 | 2026-05-19 | 28 | ernie-5.1 |
| #1 minus #10 | 2026-05-27 | 27 | gpt-5.4-high |
| #1 minus #10 | 2026-06-03 | 29 | gpt-5.4-high |
| #1 minus #10 | 2026-06-04 | 30 | gpt-5.4-high |
| #1 minus #10 | 2026-06-05 | 30 | gpt-5.4-high |
| #1 minus #10 | 2026-06-10 | 29 | muse-spark |
| #1 minus #10 | 2026-06-16 | 29 | muse-spark |
| #1 minus #10 | 2026-06-25 | 28 | muse-spark |
| #1 minus #10 | 2026-07-02 | 28 | muse-spark |
| #1 minus #10 | 2026-07-10 | 24 | gemini-3.5-flash-medium |
| #1 minus #10 | 2026-07-12 | 25 | qwen3.7-max-preview |
| #1 minus #10 | 2026-07-14 | 25 | gemini-3.5-flash-medium |
| #1 minus #10 | 2026-07-16 | 26 | qwen3.7-max-preview |
| #1 minus #10 | 2026-07-20 | 26 | qwen3.7-max-preview |
| #1 minus #10 | 2026-07-21 | 23 | gemini-3.1-pro-preview |
| #1 minus #10 | 2026-07-26 | 27 | gemini-3.1-pro-preview |
| #1 minus #10 | 2026-07-27 | 32 | gemini-3.1-pro-preview |
| #1 minus #10 | 2026-07-30 | 32 | gemini-3.1-pro-preview |
| #1 minus #10 | 2026-08-02 | 28 | gemini-3-pro |
| #1 minus #10 | 2026-08-03 | 28 | gemini-3.5-flash-high |
| #1 minus #10 | 2026-08-06 | 25 | gemini-3.5-flash-high |
| #1 minus #10 | 2026-08-10 | 26 | gemini-3.5-flash-high |
| #1 minus #10 | 2026-08-11 | 25 | gemini-3.5-flash-high |
| #1 minus #10 | 2026-08-12 | 26 | gemini-3.5-flash-high |
| #1 minus #10 | 2026-08-19 | 22 | qwen3.8-max |
| #1 minus #10 | 2026-08-21 | 22 | gemini-3.5-flash-high |
| #1 minus #10 | 2026-08-27 | 22 | claude-opus-4-7 |
| #1 minus #10 | 2026-09-01 | 22 | claude-opus-4-7 |
| #1 minus #10 | 2026-09-02 | 25 | muse-spark-1.2 (xHigh) |
| #1 minus #10 | 2026-09-11 | 21 | muse-spark-1.2 (xHigh) |
| #1 minus #10 | 2026-09-13 | 18 | muse-spark-1.3-max |
| #1 minus #20 | 2024-01-09 | 141 | zephyr-7b-beta |
| #1 minus #20 | 2024-01-18 | 141 | zephyr-7b-beta |
| #1 minus #20 | 2024-01-25 | 139 | zephyr-7b-beta |
| #1 minus #20 | 2024-02-02 | 188 | dolphin-2.2.1-mistral-7b |
| #1 minus #20 | 2024-02-15 | 174 | llama2-70b-steerlm-chat |
| #1 minus #20 | 2024-03-05 | 174 | llama2-70b-steerlm-chat |
| #1 minus #20 | 2024-03-07 | 169 | llama-2-70b-chat |
| #1 minus #20 | 2024-03-13 | 169 | llama-2-70b-chat |
| #1 minus #20 | 2024-03-26 | 166 | openchat-3.5-0106 |
| #1 minus #20 | 2024-03-27 | 165 | vicuna-33b |
| #1 minus #20 | 2024-04-03 | 157 | tulu-2-dpo-70b |
| #1 minus #20 | 2024-04-09 | 150 | yi-34b-chat |
| #1 minus #20 | 2024-04-10 | 150 | yi-34b-chat |
| #1 minus #20 | 2024-04-11 | 153 | wizardlm-70b |
| #1 minus #20 | 2024-04-13 | 152 | wizardlm-70b |
| #1 minus #20 | 2024-04-18 | 122 | gemini-pro-dev-api |
| #1 minus #20 | 2024-04-19 | 123 | gemini-pro-dev-api |
| #1 minus #20 | 2024-04-22 | 124 | qwen1.5-32b-chat |
| #1 minus #20 | 2024-04-26 | 122 | qwen1.5-32b-chat |
| #1 minus #20 | 2024-05-01 | 125 | qwen1.5-32b-chat |
| #1 minus #20 | 2024-05-08 | 122 | gemini-pro-dev-api |
| #1 minus #20 | 2024-05-15 | 145 | mixtral-8x22b-instruct-v0.1 |
| #1 minus #20 | 2024-05-16 | 143 | mixtral-8x22b-instruct-v0.1 |
| #1 minus #20 | 2024-05-19 | 142 | mixtral-8x22b-instruct-v0.1 |
| #1 minus #20 | 2024-05-20 | 142 | mixtral-8x22b-instruct-v0.1 |
| #1 minus #20 | 2024-05-27 | 133 | command-r |
| #1 minus #20 | 2024-06-02 | 130 | qwen1.5-72b-chat |
| #1 minus #20 | 2024-06-06 | 129 | mistral-large-2402 |
| #1 minus #20 | 2024-06-11 | 128 | mistral-large-2402 |
| #1 minus #20 | 2024-06-17 | 129 | mistral-large-2402 |
| #1 minus #20 | 2024-06-21 | 123 | yi-1.5-34b-chat |
| #1 minus #20 | 2024-06-23 | 118 | gpt-4-0613 |
| #1 minus #20 | 2024-06-26 | 126 | gpt-4-0613 |
| #1 minus #20 | 2024-06-29 | 108 | deepseek-coder-v2 |
| #1 minus #20 | 2024-07-06 | 109 | claude-3-haiku-20240307 |
| #1 minus #20 | 2024-07-08 | 109 | deepseek-coder-v2 |
| #1 minus #20 | 2024-07-16 | 109 | deepseek-coder-v2 |
| #1 minus #20 | 2024-07-22 | 108 | claude-3-haiku-20240307 |
| #1 minus #20 | 2024-07-25 | 108 | deepseek-coder-v2 |
| #1 minus #20 | 2024-07-30 | 98 | gemma-2-9b-it |
| #1 minus #20 | 2024-07-31 | 99 | gemma-2-9b-it |
| #1 minus #20 | 2024-08-01 | 99 | gemma-2-9b-it |
| #1 minus #20 | 2024-08-05 | 96 | command-r-plus |
| #1 minus #20 | 2024-08-06 | 96 | command-r-plus |
| #1 minus #20 | 2024-08-13 | 96 | command-r-plus |
| #1 minus #20 | 2024-08-22 | 110 | glm-4-0520 |
| #1 minus #20 | 2024-08-23 | 104 | jamba-1.5-large |
| #1 minus #20 | 2024-08-27 | 102 | jamba-1.5-large |
| #1 minus #20 | 2024-08-28 | 103 | jamba-1.5-large |
| #1 minus #20 | 2024-09-04 | 98 | gemma-2-27b-it |
| #1 minus #20 | 2024-09-15 | 74 | jamba-1.5-large |
| #1 minus #20 | 2024-09-17 | 110 | gpt-4-0125-preview |
| #1 minus #20 | 2024-09-27 | 91 | claude-3-opus-20240229 |
| #1 minus #20 | 2024-10-07 | 85 | athene-70b-0725 |
| #1 minus #20 | 2024-10-15 | 79 | deepseek-v2.5 |
| #1 minus #20 | 2024-10-23 | 80 | gpt-4-turbo-2024-04-09 |
| #1 minus #20 | 2024-10-28 | 77 | qwen2.5-72b-instruct |
| #1 minus #20 | 2024-11-04 | 75 | qwen2.5-72b-instruct |
| #1 minus #20 | 2024-11-12 | 74 | qwen2.5-72b-instruct |
| #1 minus #20 | 2024-11-13 | 74 | qwen2.5-72b-instruct |
| #1 minus #20 | 2024-11-20 | 75 | qwen2.5-72b-instruct |
| #1 minus #20 | 2024-11-21 | 75 | qwen2.5-72b-instruct |
| #1 minus #20 | 2024-11-22 | 74 | gemini-1.5-pro-001 |
| #1 minus #20 | 2024-12-01 | 74 | gemini-1.5-pro-001 |
| #1 minus #20 | 2024-12-05 | 75 | gemini-1.5-pro-001 |
| #1 minus #20 | 2024-12-10 | 74 | gemini-1.5-pro-001 |
| #1 minus #20 | 2024-12-15 | 74 | gemini-1.5-pro-001 |
| #1 minus #20 | 2024-12-18 | 75 | gemini-1.5-pro-001 |
| #1 minus #20 | 2024-12-22 | 69 | gpt-4o-2024-08-06 |
| #1 minus #20 | 2024-12-30 | 92 | llama-3.1-405b-instruct-bf16 |
| #1 minus #20 | 2025-01-05 | 85 | gemini-advanced-0514 |
| #1 minus #20 | 2025-01-15 | 83 | llama-3.1-405b-instruct-fp8 |
| #1 minus #20 | 2025-01-19 | 84 | llama-3.1-405b-instruct-fp8 |
| #1 minus #20 | 2025-01-21 | 84 | llama-3.1-405b-instruct-fp8 |
| #1 minus #20 | 2025-01-22 | 84 | llama-3.1-405b-instruct-fp8 |
| #1 minus #20 | 2025-01-24 | 89 | llama-3.1-405b-instruct-bf16 |
| #1 minus #20 | 2025-01-28 | 89 | llama-3.1-405b-instruct-bf16 |
| #1 minus #20 | 2025-02-03 | 92 | llama-3.1-nemotron-70b-instruct |
| #1 minus #20 | 2025-02-05 | 90 | gemini-1.5-flash-002 |
| #1 minus #20 | 2025-02-06 | 89 | gpt-4o-mini-2024-07-18 |
| #1 minus #20 | 2025-02-09 | 87 | glm-4-plus |
| #1 minus #20 | 2025-02-11 | 87 | glm-4-plus |
| #1 minus #20 | 2025-02-14 | 86 | athene-v2-chat |
| #1 minus #20 | 2025-02-17 | 86 | athene-v2-chat |
| #1 minus #20 | 2025-02-21 | 83 | deepseek-v2.5-1210 |
| #1 minus #20 | 2025-02-27 | 80 | qwen2.5-plus-1127 |
| #1 minus #20 | 2025-03-03 | 127 | gpt-4o-2024-05-13 |
| #1 minus #20 | 2025-03-11 | 120 | yi-lightning |
| #1 minus #20 | 2025-03-15 | 106 | claude-3-7-sonnet-20250219 |
| #1 minus #20 | 2025-03-17 | 105 | hunyuan-turbos-20250226 |
| #1 minus #20 | 2025-03-20 | 102 | gemini-1.5-pro-002 |
| #1 minus #20 | 2025-03-25 | 101 | hunyuan-turbos-20250226 |
| #1 minus #20 | 2025-03-26 | 104 | o1-mini |
| #1 minus #20 | 2025-03-30 | 101 | step-2-16k-exp-202412 |
| #1 minus #20 | 2025-04-05 | 106 | step-2-16k-exp-202412 |
| #1 minus #20 | 2025-04-08 | 106 | step-2-16k-exp-202412 |
| #1 minus #20 | 2025-04-09 | 102 | o1-mini |
| #1 minus #20 | 2025-04-16 | 103 | command-a-03-2025 |
| #1 minus #20 | 2025-04-23 | 108 | qwen-plus-0125 |
| #1 minus #20 | 2025-05-05 | 99 | gemini-2.0-flash-lite-preview-02-05 |
| #1 minus #20 | 2025-05-06 | 99 | gemini-2.0-flash-lite-preview-02-05 |
| #1 minus #20 | 2025-05-11 | 100 | qwq-32b |
| #1 minus #20 | 2025-05-19 | 91 | gemma-3-12b-it |
| #1 minus #20 | 2025-05-20 | 91 | gemma-3-12b-it |
| #1 minus #20 | 2025-05-22 | 92 | gemma-3-12b-it |
| #1 minus #20 | 2025-06-01 | 88 | o3-mini-high |
| #1 minus #20 | 2025-06-03 | 88 | o3-mini-high |
| #1 minus #20 | 2025-06-05 | 87 | o3-mini-high |
| #1 minus #20 | 2025-06-11 | 87 | o3-mini-high |
| #1 minus #20 | 2025-06-17 | 76 | o1-preview |
| #1 minus #20 | 2025-06-18 | 70 | gemma-3-27b-it |
| #1 minus #20 | 2025-06-24 | 115 | qwen2.5-max |
| #1 minus #20 | 2025-07-01 | 110 | qwen2.5-max |
| #1 minus #20 | 2025-07-07 | 114 | gemini-2.0-flash-001 |
| #1 minus #20 | 2025-07-15 | 111 | qwen2.5-max |
| #1 minus #20 | 2025-07-17 | 110 | minimax-m1 |
| #1 minus #20 | 2025-07-25 | 107 | qwen3-235b-a22b |
| #1 minus #20 | 2025-07-28 | 107 | qwen2.5-max |
| #1 minus #20 | 2025-08-01 | 103 | claude-opus-4-20250514 |
| #1 minus #20 | 2025-08-04 | 97 | deepseek-r1 |
| #1 minus #20 | 2025-08-07 | 94 | deepseek-v3-0324 |
| #1 minus #20 | 2025-08-14 | 87 | qwen3-30b-a3b-instruct-2507 |
| #1 minus #20 | 2025-08-18 | 89 | kimi-k2-0711-preview |
| #1 minus #20 | 2025-08-21 | 86 | qwen3-30b-a3b-instruct-2507 |
| #1 minus #20 | 2025-08-27 | 76 | qwen3-235b-a22b-no-thinking |
| #1 minus #20 | 2025-08-28 | 75 | qwen3-235b-a22b-no-thinking |
| #1 minus #20 | 2025-09-08 | 67 | hunyuan-t1-20250711 |
| #1 minus #20 | 2025-09-17 | 58 | gemini-2.5-flash |
| #1 minus #20 | 2025-09-18 | 58 | gemini-2.5-flash |
| #1 minus #20 | 2025-09-30 | 51 | grok-4-fast-chat |
| #1 minus #20 | 2025-10-01 | 49 | deepseek-v3.1-thinking |
| #1 minus #20 | 2025-10-03 | 48 | o3-2025-04-16 |
| #1 minus #20 | 2025-10-07 | 49 | claude-sonnet-4-5-20250929 |
| #1 minus #20 | 2025-10-16 | 48 | deepseek-v3.1-terminus-thinking |
| #1 minus #20 | 2025-10-17 | 48 | deepseek-v3.1-terminus-thinking |
| #1 minus #20 | 2025-11-05 | 48 | qwen3-235b-a22b-instruct-2507 |
| #1 minus #20 | 2025-11-06 | 48 | gpt-4.5-preview-2025-02-27 |
| #1 minus #20 | 2025-11-09 | 48 | gpt-4.5-preview-2025-02-27 |
| #1 minus #20 | 2025-11-16 | 74 | claude-opus-4-1-20250805-thinking-16k |
| #1 minus #20 | 2025-11-17 | 76 | claude-opus-4-1-20250805-thinking-16k |
| #1 minus #20 | 2025-11-18 | 73 | deepseek-v3.1 |
| #1 minus #20 | 2025-11-20 | 71 | deepseek-v3.1 |
| #1 minus #20 | 2025-11-26 | 65 | qwen3-vl-235b-a22b-instruct |
| #1 minus #20 | 2025-12-01 | 64 | longcat-flash-chat |
| #1 minus #20 | 2025-12-04 | 63 | longcat-flash-chat |
| #1 minus #20 | 2025-12-05 | 64 | longcat-flash-chat |
| #1 minus #20 | 2025-12-08 | 63 | qwen3-vl-235b-a22b-instruct |
| #1 minus #20 | 2025-12-11 | 65 | qwen3-vl-235b-a22b-instruct |
| #1 minus #20 | 2025-12-15 | 64 | qwen3-vl-235b-a22b-instruct |
| #1 minus #20 | 2025-12-16 | 64 | longcat-flash-chat |
| #1 minus #20 | 2025-12-18 | 64 | longcat-flash-chat |
| #1 minus #20 | 2025-12-19 | 62 | longcat-flash-chat |
| #1 minus #20 | 2025-12-21 | 61 | deepseek-v3.2-exp |
| #1 minus #20 | 2025-12-23 | 61 | deepseek-v3.2-exp |
| #1 minus #20 | 2025-12-30 | 60 | grok-3-preview-02-24 |
| #1 minus #20 | 2026-01-09 | 59 | gpt-5.2-high |
| #1 minus #20 | 2026-01-27 | 54 | ernie-5.0-preview-1022 |
| #1 minus #20 | 2026-01-29 | 53 | glm-4.5 |
| #1 minus #20 | 2026-02-06 | 71 | claude-sonnet-4-5-20250929 |
| #1 minus #20 | 2026-02-09 | 72 | claude-sonnet-4-5-20250929 |
| #1 minus #20 | 2026-02-10 | 70 | mistral-large-3 |
| #1 minus #20 | 2026-02-11 | 69 | glm-4.7 |
| #1 minus #20 | 2026-02-16 | 66 | qwen3-max-preview |
| #1 minus #20 | 2026-02-17 | 65 | qwen3-max-preview |
| #1 minus #20 | 2026-02-19 | 62 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-02-20 | 61 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-02-21 | 63 | gpt-5.2-chat-latest-20260210 |
| #1 minus #20 | 2026-02-22 | 63 | gpt-5.2-chat-latest-20260210 |
| #1 minus #20 | 2026-02-24 | 61 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-02-25 | 61 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-02-26 | 59 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-03-03 | 62 | ernie-5.0-preview-1203 |
| #1 minus #20 | 2026-03-04 | 61 | gpt-5.1-high |
| #1 minus #20 | 2026-03-05 | 61 | gpt-5.1-high |
| #1 minus #20 | 2026-03-06 | 61 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-03-11 | 59 | glm-5 |
| #1 minus #20 | 2026-03-16 | 54 | grok-4.1-thinking |
| #1 minus #20 | 2026-03-18 | 52 | claude-opus-4-5-20251101 |
| #1 minus #20 | 2026-03-19 | 52 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-03-20 | 52 | gpt-5.4 |
| #1 minus #20 | 2026-03-26 | 55 | grok-4.20-multi-agent-beta-0309 |
| #1 minus #20 | 2026-03-31 | 55 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-04-02 | 55 | qwen3.5-397b-a17b |
| #1 minus #20 | 2026-04-07 | 53 | claude-opus-4-5-20251101 |
| #1 minus #20 | 2026-04-09 | 54 | claude-opus-4-5-20251101-thinking-32k |
| #1 minus #20 | 2026-04-10 | 54 | claude-opus-4-5-20251101 |
| #1 minus #20 | 2026-04-14 | 53 | kimi-k2.5-thinking |
| #1 minus #20 | 2026-04-17 | 51 | gemini-3-flash (thinking-minimal) |
| #1 minus #20 | 2026-04-20 | 51 | gemini-3-flash (thinking-minimal) |
| #1 minus #20 | 2026-04-22 | 51 | amazon-nova-experimental-chat-26-02-10 |
| #1 minus #20 | 2026-04-23 | 52 | amazon-nova-experimental-chat-26-02-10 |
| #1 minus #20 | 2026-04-27 | 49 | ernie-5.0-0110 |
| #1 minus #20 | 2026-04-28 | 47 | grok-4.20-multi-agent-beta-0309 |
| #1 minus #20 | 2026-04-29 | 48 | gpt-5.4 |
| #1 minus #20 | 2026-04-30 | 48 | gpt-5.4 |
| #1 minus #20 | 2026-05-01 | 48 | gpt-5.4 |
| #1 minus #20 | 2026-05-06 | 47 | dola-seed-2.0-pro |
| #1 minus #20 | 2026-05-07 | 47 | dola-seed-2.0-pro |
| #1 minus #20 | 2026-05-12 | 48 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-05-14 | 47 | claude-sonnet-4-6 |
| #1 minus #20 | 2026-05-18 | 46 | grok-4.20-beta-0309-reasoning |
| #1 minus #20 | 2026-05-19 | 46 | gpt-5.4 |
| #1 minus #20 | 2026-05-27 | 43 | kimi-k2.6 |
| #1 minus #20 | 2026-06-03 | 43 | gemini-2.5-pro |
| #1 minus #20 | 2026-06-04 | 44 | claude-opus-4-8 |
| #1 minus #20 | 2026-06-05 | 46 | claude-opus-4-8 |
| #1 minus #20 | 2026-06-10 | 43 | gemini-2.5-pro |
| #1 minus #20 | 2026-06-16 | 39 | mimo-v2.5-pro |
| #1 minus #20 | 2026-06-25 | 38 | claude-opus-4-8-thinking |
| #1 minus #20 | 2026-07-02 | 38 | mimo-v2.5-pro |
| #1 minus #20 | 2026-07-10 | 39 | mimo-v2.5-pro |
| #1 minus #20 | 2026-07-12 | 39 | mimo-v2.5-pro |
| #1 minus #20 | 2026-07-14 | 39 | mimo-v2.5-pro |
| #1 minus #20 | 2026-07-16 | 38 | gpt-5.5 |
| #1 minus #20 | 2026-07-20 | 38 | glm-5.2 (max) |
| #1 minus #20 | 2026-07-21 | 37 | glm-5.1 |
| #1 minus #20 | 2026-07-26 | 40 | gemini-3-flash |
| #1 minus #20 | 2026-07-27 | 44 | ernie-5.1 |
| #1 minus #20 | 2026-07-30 | 44 | ernie-5.1 |
| #1 minus #20 | 2026-08-02 | 38 | gpt-5.4-high |
| #1 minus #20 | 2026-08-03 | 40 | ernie-5.1 |
| #1 minus #20 | 2026-08-06 | 35 | gpt-5.5-high |
| #1 minus #20 | 2026-08-10 | 38 | gpt-5.5-high |
| #1 minus #20 | 2026-08-11 | 37 | gpt-5.5-high |
| #1 minus #20 | 2026-08-12 | 37 | gpt-5.5-high |
| #1 minus #20 | 2026-08-19 | 32 | muse-spark |
| #1 minus #20 | 2026-08-21 | 31 | muse-spark |
| #1 minus #20 | 2026-08-27 | 31 | muse-spark |
| #1 minus #20 | 2026-09-01 | 31 | muse-spark |
| #1 minus #20 | 2026-09-02 | 39 | qwen3.7-max-preview |
| #1 minus #20 | 2026-09-11 | 36 | qwen3.7-max-preview |
| #1 minus #20 | 2026-09-13 | 32 | glm-5.3-max |

Note: Ranks count every listed model, so effort variants of one model can fill several places: the 13 Sept 2026 top 10 holds models from only 3 labs. Counting each lab once, the gap from #1 to the fifth-best lab fell from 84 points (Jan 2024) to 32. Win rate uses Arena's scale: P(lower wins) = 1 / (1 + 10^(gap/400)).

## The signal

- **Launches:** DeepSeek V4.1 Flash (10 Sep) and Kimi K2.8 Preview (11 Sep) were announced on consecutive days.
- **Arena:** 13 Sept 2026: muse-spark-1.3-max (Meta) entered the top 10 for the first time, at #10.
- **Arena:** On the same snapshot, 6 models have a confidence interval that reaches claude-fable-5.1-max's lower bound, so they are statistically tied with #1.

## The Mode read

In the days before this snapshot, two more labs shipped models: DeepSeek with DeepSeek V4.1 Flash and Moonshot with Kimi K2.8 Preview. On the 13 Sept 2026 Arena snapshot, claude-fable-5.1-max is #1 and muse-spark-1.3-max is #10. The question behind every "LLM leaderboard top 10" search is how much the ranking matters.

Less and less. In Feb 2024 the #1 model led #10 by **148 points**; today the gap is **18**, and #1 to #20 is 32 (it was 188 in Feb 2024). At 18 points, #10 wins about 47% of blind votes against #1. On this snapshot 6 models are inside #1's confidence interval. Some of this is one lab's variants clustering, but counting each lab once the gap still fell from 84 to 32 points.

When ten models are this close, a buyer picks on other things: price, speed, data rules, and which vendor is already on the purchase order. That last one is distribution. A model sold inside a cloud account, a SaaS seat or an integrator's contract gets chosen without a bake-off, and the ranking rarely gets checked again.

> When the top ten are tied, the buyer picks the vendor already on the invoice.

**What we would back:** Companies that sell AI through a channel the buyer already pays: a cloud marketplace listing, an existing SaaS seat, an integrator.

## Sources

- [Arena leaderboard dataset (lmarena-ai, CC BY 4.0)](https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset)
- [Launch feed (lab announcements, captured 7 Oct 2026)](undefined)
