Artificial Analysis Coding (September 2026)
In September 2026, GPT-5.6 Sol (xhigh) ranks first with 78.3, followed by GPT-5.6 Sol (max) (77.4) and GPT-5.6 Sol (high) (77.2). Top 20 of 446 candidates.
Data as of September 2026
#ModelProviderScoreIntelligence
1GGPT-5.6 Sol (xhigh)OpenAI78.359.0
2GGPT-5.6 Sol (max)OpenAI77.460.9
3GGPT-5.6 Sol (high)OpenAI77.257.3
4GGPT-5.6 Terra (max)OpenAI76.756.6
5CClaude Fable 5Anthropic76.562.1
6GGPT-5.6 Sol (medium)OpenAI76.355.6
7KKimi K3 (max)Kimi76.259.7
8GGPT-5.5OpenAI74.956.3
9CClaude Opus 4.8 (Adaptive Reasoning, Max Effort)Anthropic74.357.3
10CClaude Opus 4.7Anthropic73.655.0
11GGrok 4.5 (high)SpaceXAI72.455.8
12CClaude Sonnet 5Anthropic71.555.3
13GGPT-5.6 Luna (max)OpenAI71.452.3
14MMuse Spark 1.1 (xhigh)Meta71.353.2
15GGPT-5.4OpenAI71.153.1
16GGPT-5.6 Terra (xhigh)OpenAI70.652.8
17GGemini 3.5 Flash (high)Google70.152.0
18GGPT-5.6 Sol (low)OpenAI69.750.7
19GGemini 3.1 ProGoogle68.847.7
20GGLM-5.2Z AI68.852.6
How this chart is built
Public scores from a third-party model evaluation service. The source replaces its full table each release and keeps no history, so no trend or month-over-month change is shown.
Positions are renumbered from high to low by this table's own metric, independent of any ranking the source publishes. Figures are the current real data with no manual adjustment; values we cannot obtain show as “—” and are never filled in with estimates.