← the full leaderboard · coding
The best AI model for coding, ranked.
Gemini 4 Carbon leads the coding ranking today, ahead of Gemini 4 Argon and Claude Opus 5.5. LMArena · Coding, LiveBench · Coding and Vals Vibe Code Bench: 47 models combined into one coding index, updated daily. Free and public.
Scores as published by each board; table composed November 11, 2026.
Get the next launch alert, free →
Coding leaderboard
| # | Model | Coding index | LMArena · Coding | LiveBench · Coding | Vals Vibe Code Bench | API price / 1M |
|---|---|---|---|---|---|---|
| 1 | Gemini 4 Carbon · Google | 100 | 1,563 | 91.0 | 89.1 | $2 / $10 |
| 2 | Gemini 4 Argon · Google | 95.2 | 1,540 | 85.7 | 91.9 | $2 / $10 |
| 3 | Claude Opus 5.5 · Anthropic | 89.8 | 1,542 | 87.1 | 90.3 | $4 / $20 |
| 4 | Claude Fable 5 · Anthropic | 88.8 | 1,519 | — | 90.4 | $10 / $50 |
| 5 | Claude Fable 5.1 · Anthropic | 88.1 | 1,527 | 89.7 | 90.3 | $10 / $50 |
| 6 | GPT-6 Astra · OpenAI | 85.7 | 1,494 | 87.5 | 89.6 | $10 / $50 |
| 7 | Claude Opus 5 · Anthropic | 85 | 1,532 | — | 88.4 | $5 / $25 |
| 8 | Claude Fable 5.2 · Anthropic | 81.5 | — | 79.2 | 93.8 | $8 / $40 |
| 9 | GPT-6.1 Terra · OpenAI | 76.9 | — | 89.2 | 84.0 | $1 / $5 |
| 10 | Muse Spark 1.3 · Meta | 74.1 | 1,508 | 83.1 | 85.9 | $1.25 / $4.25 |
| 11 | DeepSeek V4.1 Flash · DeepSeek | 71.4 | 1,503 | 80.4 | 87.7 | — |
| 11 | Muse Wonder · Meta | 71.4 | 1,517 | 74.9 | 86.0 | — |
| 13 | GPT-6.1 Sol · OpenAI | 70.4 | 1,486 | 82.1 | 88.9 | $2 / $10 |
| 14 | MiMo-V2.6 Pro · Xiaomi | 67.3 | 1,533 | 64.5 | 85.2 | $0.435 / $0.87 |
| 15 | Claude Sonnet 5.5 · Anthropic | 66.7 | 1,500 | 77.2 | 92.4 | $2 / $10 |
| 16 | Kimi K3 · Moonshot | 65.3 | 1,505 | 53.1 | 85.0 | $3 / $15 |
| 17 | Claude Haiku 5.5 · Anthropic | 59.3 | 1,454 | 77.6 | 90.4 | $0.10 / $0.50 |
| 18 | GPT-5.6 · OpenAI | 57.1 | 1,495 | 84.2 | 80.5 | $4 / $20 |
| 19 | Claude Opus 4.8 · Anthropic | 52.4 | 1,491 | — | 82.7 | $5 / $25 |
| 20 | GLM-5.3 · Zhipu AI | 51.5 | 1,499 | — | 78.1 | $1.40 / $4.40 |
| 21 | Muse Spark 1.1 · Meta | 45.2 | 1,498 | — | 72.2 | $1.25 / $4.25 |
| 22 | Grok 4.7 · xAI | 44.4 | 1,425 | 73.1 | 86.2 | $2 / $6 |
| 23 | Claude Sonnet 5 · Anthropic | 42.5 | 1,483 | — | 81.3 | $2 / $10 |
| 24 | GPT-6 Sol · OpenAI | 41.2 | 1,448 | — | 87.8 | $2 / $10 |
| 25 | Gemini 3.8 Flash · Google | 40.8 | 1,511 | 65.7 | 78.7 | $0.75 / $3.75 |
| 26 | GLM-5.3 Flash · Zhipu AI | 40.7 | 1,510 | 70.9 | 30.8 | $0.15 / $0.50 |
| 27 | DeepSeek V4 Pro · DeepSeek | 39.5 | 1,469 | — | 82.3 | $0.66 / $1.98 |
| 28 | GPT-5.5 · OpenAI | 38.4 | 1,494 | — | 69.8 | $5 / $30 |
| 29 | GPT-6 Luna · OpenAI | 37 | 1,432 | 70.8 | 81.6 | $0.10 / $0.50 |
| 30 | Qwen 3.7 Max · Alibaba | 36.9 | 1,498 | — | 47.7 | — |
| 31 | Qwen 3.8 Max · Alibaba | 36.7 | 1,503 | 60.9 | 77.1 | $2 / $6 |
| 32 | Grok 4.6 · xAI | 34.7 | 1,464 | 76.7 | 76.2 | $2 / $6 |
| 33 | Step 5 Preview · StepFun | 33.3 | 1,482 | 69.4 | 70.4 | — |
| 34 | Gemini 3.5 Flash · Google | 33.2 | 1,493 | — | 48.7 | $1.50 / $9 |
| 35 | Kimi K3.1 Pro · Moonshot | 31 | 1,477 | 67.8 | 84.9 | $2 / $10 |
| 36 | GLM-5.2 · Zhipu AI | 30.6 | 1,488 | — | 64.0 | $1.40 / $4.40 |
| 37 | Inkling 2 Small · Thinking Machines | 30.4 | 1,457 | — | 81.5 | — |
| 38 | MiniMax M3.1 Flash · MiniMax | 30.1 | — | 68.8 | 73.9 | $0.15 / $0.60 |
| 39 | Kimi K2.6 · Moonshot | 26.7 | 1,489 | — | 37.9 | — |
| 40 | Beam · Reflection AI | 24.5 | 1,475 | — | 66.4 | — |
| 41 | Grok 4.5 · xAI | 24.3 | 1,474 | — | 69.0 | $2 / $6 |
| 42 | Muse Spark Open · Meta | 21.9 | — | 55.1 | 76.2 | — |
| 43 | Gemini 3.1 Pro · Google | 20.9 | 1,482 | — | 32.0 | — |
| 44 | MiniMax M3 · MiniMax | 16.8 | 1,469 | — | 47.6 | $0.30 / $1.20 |
| 45 | Mistral Large 4 · Mistral | 14.3 | 1,457 | 44.4 | 81.6 | $0.70 / $2.10 |
| 46 | Inkling · Thinking Machines | 3.7 | 1,467 | 48.7 | 19.2 | — |
| 47 | Inkling Small · Thinking Machines | 3.6 | 1,451 | — | 19.1 | — |
| — | Muse Ember · Meta | — | — | — | 83.7 | — |
| — | Hy4 · Tencent | — | — | — | 81.0 | — |
| — | Kimi K2.7 Code · Moonshot | — | — | — | 47.2 | — |
Coding index: the median of each model's percentile across the boards that score it (at least two). Prices are input / output per 1M tokens on the maker's own API. A dash means that board doesn't score the model, or the maker publishes no API price.
Quick answers
What is the best AI model for coding right now?
Gemini 4 Carbon at the top of the ArtificialWatch coding index (100), ahead of Gemini 4 Argon (95.2) and Claude Opus 5.5 (89.8). The index is the median of each model's percentile across LMArena · Coding, LiveBench · Coding and Vals Vibe Code Bench, so no single board decides it; each board's own score is in the table.
What is the best open-weights model for coding?
DeepSeek V4.1 Flash from DeepSeek, the highest-ranked model with open weights on the coding index (71.4, #11 overall).
What is the cheapest top-10 AI model for coding?
Among the ten highest-ranked models with a price on their maker's own API, Muse Spark 1.3 (#10) has the lowest output price: $1.25 input / $4.25 output per 1M tokens.
Which benchmarks does the coding ranking combine?
LMArena · Coding (Elo (community votes on coding prompts)); LiveBench · Coding (mean of its Coding and Agentic Coding categories (0-100)); Vals Vibe Code Bench (accuracy %). Each model's best-variant score on each board becomes a percentile within that board, and a model needs at least 2 boards to get an index score. Vals' SWE-bench Verified, LiveCodeBench and Terminal-Bench are left out: Vals archived them and no longer runs new models on them.
The data
This table is free to use: JSON · CSV, rebuilt every day. The index and the percentile ranks are ours, under CC BY 4.0; each board's raw scores belong to that board, and the JSON names and links every source. All our open data.
Cite asArtificialWatch Coding Index, 2026-11-11. ArtificialWatch. https://artificialwatch.com/coding
Sources
- LMArena · Coding · Elo (community votes on coding prompts) · 43 models · read November 11, 2026
- LiveBench · Coding · mean of its Coding and Agentic Coding categories (0-100) · 28 models · read November 11, 2026
- Vals Vibe Code Bench · accuracy % · 50 models · read November 11, 2026
The full LLM leaderboard · Open-weights models ranked · Every model's API price · Head-to-head comparisons
Use this chart
Writing about AI models for coding? Paste this where you want the chart. It redraws when the table updates, and the credit line links back to this page.
<figure><a href="https://artificialwatch.com/coding"><img src="https://artificialwatch.com/api/aw/og/leaderboard/coding" alt="The top five AI models for coding on the ArtificialWatch coding index" width="600" height="315"></a><figcaption>Chart: <a href="https://artificialwatch.com/coding">ArtificialWatch</a>, CC BY 4.0</figcaption></figure>Reports from this data
Every report, from every dataset: data reports.
Get an email when a new model takes #1. Free.
One email when this leaderboard or the bench index gets a new #1 (about once a month so far), plus alerts when a state-of-the-art model launches. Change either in your settings.
- Free forever
- No card
- Unsubscribe in one click