Best AI Models for Coding
4 models ranked by coding benchmark performance across HumanEval, SWE-bench, and other coding evaluations.
Best Overall
Claude Opus 4.5 (batch)
Anthropic · Avg: 80.9
Best Value for Coding
GPT-5.2
OpenAI · $0.88/M in
| # | Model | Coding Avg | Price |
|---|---|---|---|
| 1 | Claude Opus 4.5 (batch) Anthropic | 80.9 | $2.50 |
| 2 | Claude Opus 4.6 (batch) Anthropic | 80.8 | $5.00 |
| 3 | GPT-5.2 OpenAI | 80.0 | $0.88 |
| 4 | Claude Sonnet 4.5 (batch) Anthropic | 77.2 | $1.50 |
Other Notable Models
These models don't have published coding benchmark scores yet but are commonly used for coding tasks.
Claude Sonnet 4.6 (batch)
Anthropic · Awaiting coding benchmark evidence
GPT-5.2 Pro (batch)
OpenAI · Awaiting coding benchmark evidence
Claude Haiku 4.5 (batch)
Anthropic · Awaiting coding benchmark evidence
GPT-5 Pro (batch)
OpenAI · Awaiting coding benchmark evidence
DeepSeek V3.2
DeepSeek · Awaiting coding benchmark evidence
GPT-5 (batch)
OpenAI · Awaiting coding benchmark evidence
GPT-5 Nano (batch)
OpenAI · Awaiting coding benchmark evidence
O3 Pro
OpenAI · Awaiting coding benchmark evidence
Mistral Large
Mistral · Awaiting coding benchmark evidence
Reka Flash 3
Reka · Awaiting coding benchmark evidence