LLMs TigerLive LLM benchmarks

Aider Code Editing leaderboard: which LLMs score highest

Aider Code Editing (Aider official benchmark) currently has scores for 92 models. The top score is 84.2% by Claude 3.5-Sonnet-20241022; the median model scores 55.6% and the lowest scores 14.3%.

Aider Code Editing leaderboard: top 50 of 92 models (highest-effort row per model)
#ModelOrganisationAider Code Editing scoreCheapest input $/MSourceCaptured
1Claude 3.5-Sonnet-20241022Anthropic84.2%$3.00Aider Aider Code Editing leaderboard2026-10-05
2gemini-exp-1206Google DeepMind,Google80.5%$0.300Aider Aider Code Editing leaderboard2026-10-05
3o1-previewOpenAI79.7%$15.00Aider Aider Code Editing leaderboard2026-10-05
4claude-3.5-sonnet-20240620Anthropic77.4%$3.00Aider Aider Code Editing leaderboard2026-10-05
5claude-3-5-haiku-20241022Anthropic75.2%$1.00Aider Aider Code Editing leaderboard2026-10-05
6DeepSeek Coder V2 0724DeepSeek72.9%—Aider Aider Code Editing leaderboard2026-10-05
7gpt-4o-2024-05-13OpenAI72.9%$5.00Aider Aider Code Editing leaderboard2026-10-05
8ollama/qwen2.5-coder:32bAlibaba72.9%—Aider Aider Code Editing leaderboard2026-10-05
9DeepSeek-V2.5deepseek-ai72.2%—Aider Aider Code Editing leaderboard2026-10-05
10DeepSeek-V2.5-1210DeepSeek72.2%—Aider Aider Code Editing leaderboard2026-10-05
11openai/chatgpt-4o-latestOpenAI72.2%—Aider Aider Code Editing leaderboard2026-10-05
12gpt-4o-2024-08-06OpenAI71.4%$2.50Aider Aider Code Editing leaderboard2026-10-05
13gpt-4o-2024-11-20OpenAI71.4%$2.50Aider Aider Code Editing leaderboard2026-10-05
14Qwen2.5 Coder 32B InstructQwen71.4%$0.060Aider Aider Code Editing leaderboard2026-10-05
15DeepSeek Chat V2 0628DeepSeek69.9%—Aider Aider Code Editing leaderboard2026-10-05
16gemini-2.0-flash-expGoogle DeepMind,Google69.9%—Aider Aider Code Editing leaderboard2026-10-05
17Qwen2.5-Coder-14B-InstructAlibaba69.2%—Aider Aider Code Editing leaderboard2026-10-05
18claude-3-opus-20240229Anthropic68.4%$15.00Aider Aider Code Editing leaderboard2026-10-05
19GPT-4 (0613)OpenAI67.7%$30.00Aider Aider Code Editing leaderboard2026-10-05
20Dracarys2-72B-Instruct—66.9%—Aider Aider Code Editing leaderboard2026-10-05
21gemini-1.5-pro-exp-0827Google66.9%—Aider Aider Code Editing leaderboard2026-10-05
22gpt-4-0125-previewOpenAI66.2%$10.00Aider Aider Code Editing leaderboard2026-10-05
23gpt-4-0314OpenAI66.2%—Aider Aider Code Editing leaderboard2026-10-05
24Llama-3.1-405B-InstructMeta AI66.2%$3.50Aider Aider Code Editing leaderboard2026-10-05
25gemini-1.5-pro-002Google DeepMind65.4%—Aider Aider Code Editing leaderboard2026-10-05
26gpt-4-1106-previewOpenAI65.4%$10.00Aider Aider Code Editing leaderboard2026-10-05
27mistral-large-2411Mistral AI65.4%$2.00Aider Aider Code Editing leaderboard2026-10-05
28openrouter/qwen/qwen-2.5-coder-32b-instructAlibaba65.4%—Aider Aider Code Editing leaderboard2026-10-05
29qwen-2.5-72b-instructAlibaba65.4%$0.380Aider Aider Code Editing leaderboard2026-10-05
30yi-lightning—65.4%—Aider Aider Code Editing leaderboard2026-10-05
31gpt-4-turbo-2024-04-09OpenAI63.9%$10.00Aider Aider Code Editing leaderboard2026-10-05
32nousresearch/hermes-3-llama-3.1-405b—63.9%$1.00Aider Aider Code Editing leaderboard2026-10-05
33ollama/Qwen2.5.1-Coder-7B-Instruct-GGUF:Q8_0-32kAlibaba63.9%—Aider Aider Code Editing leaderboard2026-10-05
34ollama/qwen2.5-coder:14bAlibaba61.7%—Aider Aider Code Editing leaderboard2026-10-05
35gemini-exp-1114Google60.9%—Aider Aider Code Editing leaderboard2026-10-05
36Mistral Large 2 (2407)Mistral AI60.2%—Aider Aider Code Editing leaderboard2026-10-05
37Llama-3.3-70B-InstructMeta AI59.4%$0.120Aider Aider Code Editing leaderboard2026-10-05
38Grok-2xAI58.6%$2.00Aider Aider Code Editing leaderboard2026-10-05
39Llama-3.1-70B-InstructMeta AI58.6%$0.400Aider Aider Code Editing leaderboard2026-10-05
40ollama/qwen2.5:32b-instruct-q8_0Alibaba58.6%—Aider Aider Code Editing leaderboard2026-10-05
41gemini-exp-1121Google57.9%—Aider Aider Code Editing leaderboard2026-10-05
42gpt-3.5-turbo-0301OpenAI57.9%—Aider Aider Code Editing leaderboard2026-10-05
43Qwen2.5-Coder-7B-InstructAlibaba57.9%$0.010Aider Aider Code Editing leaderboard2026-10-05
44gemini-1.5-pro-001Google DeepMind57.1%—Aider Aider Code Editing leaderboard2026-10-05
45gpt-3.5-turbo-1106OpenAI56.1%$1.00Aider Aider Code Editing leaderboard2026-10-05
46GPT-4o miniOpenAI55.6%$0.150Aider Aider Code Editing leaderboard2026-10-05
47qwen2-72b-instructAlibaba55.6%$0.900Aider Aider Code Editing leaderboard2026-10-05
48claude-3-sonnet-20240229Anthropic54.9%$3.00Aider Aider Code Editing leaderboard2026-10-05
49Grok-2-minixAI54.9%—Aider Aider Code Editing leaderboard2026-10-05
50Llama-3.1-Nemotron-70B-Instruct-HFNVIDIA54.9%$0.880Aider Aider Code Editing leaderboard2026-10-05

Compare all benchmarks side by side on the live leaderboard →

Among the ten highest scorers, the cheapest listed API price belongs to gemini-exp-1206 at $0.300 per million input tokens (score 80.5%).

What Aider Code Editing measures

Aider's code-editing benchmark measures how reliably a model can change existing code by emitting edits that the Aider tool can apply, then running the exercise tests.

How to read these scores

Scores are percentages (higher is better): the share of questions or tasks the model completed correctly. Where a model is listed at several reasoning efforts, only its highest-effort row is shown. See the methodology for how rows are chosen.

Where the data comes from

Aider Aider Code Editing leaderboard. Scores were last captured 2026-10-05; the Captured column gives each row's date.

Related benchmarks