Aider polyglot leaderboard
- GPT-5 leads
- 88.0%
- models measured
- 22
- lead over #2
- +3.1 pts
- median
- 55.8%
Leader on Aider polyglot · read October 22
# Aider polyglot, read October 22
leaders = {
"GPT-5": 88.0%, "o3 Pro": 84.9%, "Gemini 2.5 Pro Preview 06-05": 83.1%,
}
On October 22, 2026, 29 of our catalog's models had a Aider polyglot result, and OpenAI GPT-5 led with 88.0%, ahead of OpenAI o3 Pro at 84.9%.
Aider gives a model coding exercises in C++, Go, Java, JavaScript, Python and Rust and checks whether the edited code passes the exercise tests. The number is the share of exercises solved.
Role in the index
A coding board we show beside the Vibe Coding Index, not inside it: the index keeps the six boards it was defined on, so a score does not move every time a board is added. Read the methodology
| # | Model | Result | Cost of the run | Solved on the first try | Edits well formed | Time per exercise | Input tokens, whole run | Output tokens, whole run | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | GPT-5OpenAI | 88.0% | $29.08 | 52.0% | 91.6% | 3m 14s | 2.7M | 2.6M | 38.8 | $10.00 | |
| 2 | o3 ProOpenAI | 84.9% | $146.32 | 43.6% | 97.8% | 7m 29s | 2.4M | 1.2M | Not benchmarked | $80.00 | |
| 3 | Gemini 2.5 Pro Preview 06-05Google | 83.1% | $49.88 | 46.2% | 99.6% | 3m 20s | 2.7M | 4.6M | Not benchmarked | $10.00 | |
| 4 | o3OpenAI | 81.3% | $21.23 | 40.0% | 94.7% | 3m 17s | 3.1M | 2.0M | Not benchmarked | $8.00 | |
| 5 | DeepSeek V3.2 ExpDeepSeek | 74.2% | $1.30 | 39.6% | 97.3% | 4m 51s | 2.2M | 1.6M | Not benchmarked | $0.41 | |
| 6 | Claude Opus 4RetiredAnthropic | 72.0% | $65.75 | 37.3% | 97.3% | 44 s | 2.6M | 363,142 | Not benchmarked | $75.00 | |
| 7 | o4 MiniOpenAI | 72.0% | $19.64 | 19.6% | 90.7% | 2m 57s | Not published | Not published | Not benchmarked | $4.40 | |
| 8 | R1 0528DeepSeek | 71.4% | $4.80 | 34.4% | 94.6% | 11m 57s | 2.6M | 1.8M | Not benchmarked | $2.18 | |
| 9 | DeepSeek V3DeepSeek | 70.2% | $0.88 | 38.7% | 98.2% | 1m 44s | 2.3M | 573,477 | Not benchmarked | $0.9594 | |
| 10 | o1OpenAI | 61.7% | $186.50 | 23.7% | 91.5% | 2m 13s | Not published | Not published | Not benchmarked | $60.00 | |
| 11 | Claude 3.7 SonnetRetiredAnthropic | 60.4% | $17.72 | 24.4% | 93.3% | 28 s | Not published | Not published | Not benchmarked | $15.00 | |
| 12 | o3 MiniOpenAI | 60.4% | $18.16 | 21.0% | 93.3% | 2m 05s | Not published | Not published | Not benchmarked | $4.40 | |
| 13 | Kimi K2 0711Moonshot AI | 59.1% | $1.24 | 20.4% | 92.9% | 1m 08s | 2.4M | 363,846 | Not benchmarked | $2.30 | |
| 14 | R1RetiredDeepSeek | 56.9% | $5.42 | 26.7% | 96.9% | 1m 54s | Not published | Not published | Not benchmarked | $2.50 | |
| 15 | Grok 3RetiredxAI | 53.3% | $11.03 | 22.2% | 99.6% | 15 s | Not published | Not published | Not benchmarked | $15.00 | |
| 16 | GPT-4.1OpenAI | 52.4% | $9.86 | 20.0% | 98.2% | 21 s | Not published | Not published | Not benchmarked | $8.00 | |
| 17 | Claude 3.5 SonnetRetiredAnthropic | 51.6% | $14.41 | 22.2% | 99.6% | 21 s | Not published | Not published | Not benchmarked | $15.00 | |
| 18 | gpt-oss-120bOpenAI | 41.8% | $0.74 | 13.8% | 79.1% | 36 s | 3.1M | 856,495 | Not benchmarked | $0.50 | |
| 19 | Qwen3 32BQwen | 40.0% | $0.76 | 14.2% | 83.6% | 6m 12s | 317,591 | 120,418 | Not benchmarked | $0.40 | |
| 20 | o1-miniRetiredOpenAI | 32.9% | $18.58 | 5.8% | 96.9% | 35 s | Not published | Not published | Not benchmarked | $4.40 | |
| 21 | GPT-4.1 MiniOpenAI | 32.4% | $1.99 | 11.1% | 92.4% | 20 s | Not published | Not published | Not benchmarked | $1.60 | |
| 22 | Claude 3.5 HaikuRetiredAnthropic | 28.0% | $6.06 | 7.1% | 91.1% | 32 s | Not published | Not published | Not benchmarked | $4.00 | |
| 23 | GPT-4o (2024-08-06)OpenAI | 23.1% | $7.03 | 4.9% | 94.2% | 16 s | Not published | Not published | Not benchmarked | $10.00 | |
| 24 | GPT-4o (2024-11-20)OpenAI | 18.2% | $6.74 | 4.9% | 95.1% | 12 s | Not published | Not published | Not benchmarked | $10.00 | |
| 25 | Qwen2.5 Coder 32B InstructQwen | 16.4% | Not published | 4.9% | 99.6% | 42 s | Not published | Not published | Not benchmarked | $1.00 | |
| 26 | Llama 4 MaverickMeta Llama | 15.6% | Not published | 4.4% | 99.1% | 21 s | Not published | Not published | Not benchmarked | $1.00 | |
| 27 | GPT-4.1 NanoOpenAI | 8.9% | $0.43 | 3.1% | 94.2% | 12 s | Not published | Not published | Not benchmarked | $0.40 | |
| 28 | Gemma 3 27BGoogle | 4.9% | Not published | 1.8% | 100.0% | 1m 20s | Not published | Not published | Not benchmarked | $0.30 | |
| 29 | GPT-4o-mini (2024-07-18)OpenAI | 3.6% | $0.32 | 0.9% | 100.0% | 17 s | Not published | Not published | Not benchmarked | $0.60 |
Only models in our catalog are listed, and each appears once with its best published run. 29 rows were read from the source on October 22, 2026, and rows that match no catalog model are left out. Cost of the run is what Aider reports the whole run of 225 exercises cost, in US dollars. Every other column is a figure the board publishes for the same run.