Coding AI model rankings
The Coding track ranks 40 AI models by a 50/50 blend of SWE-Bench Verified and LiveCodeBench v6. Snapshot 2026-10-14, updated weekly.
Snapshot 2026-10-14 · updated weekly
| Rank | Model | Vendor | SWE-Bench | GPQA | Context | Price (in / out) | Speed |
|---|---|---|---|---|---|---|---|
1— | Claude Opus 5.5 | Anthropic | 82.8 | 95.8% | 500K | $8 / $40 | 88 |
2▲1 | GPT-6.1 Sol | OpenAI | 81.8 | 94.9% | 1M | $6 / $36 | 112 |
3▼1 | GPT-6 Astra | OpenAI | 80.8 | 96% | 1M | $5 / $30 | 128 |
4— | Claude Fable 5.1 | Anthropic | 79.6 | 93.4% | 400K | $4 / $20 | 142 |
5— | Claude Sonnet 5.5 | Anthropic | 78.8 | 92.8% | 500K | $3 / $15 | 156 |
6— | Claude Fable 5 | Anthropic | 77.6 | 91.9% | 400K | $3.50 / $18 | 138 |
7▼1 | Claude Opus 5 | Anthropic | 77.3 | 90.7% | 400K | $6 / $30 | 96 |
8▲2 | Gemini 4 Argon | 76.8 | 92.1% | 2M | $2.50 / $12 | 165 | |
9— | GPT-6 Sol | OpenAI | 75.3 | 90.2% | 800K | $4 / $24 | 118 |
10NEW | DeepSeek Harness | DeepSeek | 74.4 | 86.5% | 256K | $0.35 / $1.40 | 105 |
11— | GPT-5.6 Sol | OpenAI | 73.2 | 88.4% | 512K | $2.80 / $16 | 134 |
12▲1 | MiniMax H3 | MiniMax | 72.5 | 85.8% | 1M | $0.60 / $2.20 | 121 |
13— | Gemini 4.0 | 72.1 | 87.9% | 2M | $2.20 / $10 | 158 | |
14— | Kimi K3 | Moonshot AI | 70.8 | 93.5% | 512K | $0.50 / $2 | 98 |
15— | Muse Wonder Full | Meta | 69.5 | 88.1% | 512K | $0.90 / $3.20 | 88 |
16NEW | Claude Haiku 5.5 | Anthropic | 69.1 | 83.2% | 500K | $0.80 / $4 | 232 |
17— | GLM-5.3 | Zhipu AI | 68.3 | 85.2% | 256K | $0.40 / $1.50 | 118 |
18— | DeepSeek V4.5 | DeepSeek | 67.6 | 83.7% | 256K | $0.30 / $1.20 | 112 |
19— | Muse Wonder | Meta | 67.5 | 85.9% | 256K | $0.45 / $1.60 | 126 |
20▲1 | Qwen3.8 Next | Alibaba | 66.5 | 82.9% | 512K | $0.50 / $1.80 | 104 |
21— | Grok-4 Fast Reasoning | xAI | 64.3 | 81.5% | 2M | $0.80 / $3.20 | 210 |
22▲3 | GLM-5.3 Flash | Zhipu AI | 63.7 | 80.4% | 256K | $0.12 / $0.50 | 268 |
23— | Mistral Large 3.5 | Mistral AI | 63.3 | 80.7% | 256K | $0.80 / $2.40 | 132 |
24— | Gemini 3 Pro | 62.6 | 80.9% | 1M | $1.50 / $6 | 168 | |
25— | Gemini 3.1 Pro | 61.3 | 79.8% | 1M | $1.25 / $5 | 172 | |
26▲2 | Gemini 3.8 Flash | 59.2 | 78.9% | 1M | $0.15 / $0.60 | 342 | |
27— | Qwen3.8 27B | Alibaba | 57.6 | 76.8% | 256K | $0.15 / $0.60 | 186 |
28— | MiMo-V2.6-Pro | Xiaomi | 56.2 | 75.4% | 256K | $0.25 / $1 | 142 |
29— | GPT-5.6 Luna | OpenAI | 55.8 | 74.2% | 512K | $1.20 / $5.50 | 286 |
30— | MiniMax H3 Flash | MiniMax | 53.6 | 72.6% | 512K | $0.10 / $0.40 | 358 |
31— | Kimi K3 Flash | Moonshot AI | 50.5 | 70.2% | 256K | $0.14 / $0.55 | 296 |
32— | Qwen3.8 Flash | Alibaba | 46.7 | 66.8% | 256K | $0.06 / $0.25 | 412 |
33— | Gemma 4 27B | 45.3 | 67.4% | 256K | $0.18 / $0.70 | 174 | |
34NEW | Strata 2 | Strata Labs | 41.4 | 62.8% | 128K | $0.12 / $0.45 | 248 |
35NEW | Jev 1 | Jev Labs | 38.0 | 59.4% | 128K | $0.08 / $0.30 | 322 |
36NEW | Clef 2 | Clef AI | 35.8 | 56.9% | 128K | $0.10 / $0.35 | 276 |
37— | Mercury 2.5 | Inception Labs | 31.4 | 48.2% | 128K | $0.10 / $0.40 | 583 |
38— | Celeris-1 | Celeris | 25.1 | 42.6% | 128K | $0.20 / $0.80 | 1431 |
39NEW | Laya 1 | Laya Research | 21.3 | 38.4% | 64K | $0.05 / $0.20 | 368 |
40NEW | Pixel Canary | 13.8 | 31.2% | 32K | Free / Free | 210 |
Coding FAQ
Which AI model is best for Coding?
Claude Opus 5.5 leads the Coding track as of October 2026. The full table above ranks all 40 tracked models on this dimension, with the underlying values shown for every row.
How is the Coding ranking calculated?
This track ranks models by a 50/50 blend of SWE-Bench Verified and LiveCodeBench v6. Scores are re-verified weekly against public benchmark sources and API rate cards, and the snapshot date is shown on every table so you can see exactly how fresh each value is.