Best for Coding

378 models

Updated 50m ago

AI models ranked by coding ability across benchmarks, real-world usage, and developer sentiment.

	Model	Provider				14d	Status
#1	A Claude Fable 5Anthropic	A Anthropic	97.1	-	-
#2	A Claude Fable 5 (batch)Anthropic	A Anthropic	97.1	-	377	-	New Hot
#3	A Claude Opus 5 (Fast)Anthropic	A Anthropic	95.1	-	171		May change Hot
#4	A Claude Opus 5Anthropic	A Anthropic	95.1	-	171		May change Hot
#5	A Claude Opus 4.8 (Fast)Anthropic	A Anthropic	95.1	-	1		Falling
#6	A Claude Opus 4.8Anthropic	A Anthropic	95.1	-	1		Falling
#7	A Claude Opus 4.7 (Fast)Anthropic	A Anthropic	95.1	-	5		Falling
#8	A Claude Opus 4.7Anthropic	A Anthropic	95.1	-	5		Falling
#9	A Claude Opus 4.7 (batch)Anthropic	A Anthropic	95.1	-	370	-	New Hot
#10	A Claude Opus 4.8 (batch)Anthropic	A Anthropic	94.6	-	369	-	New Hot
#11	O GPT-5.5 ProOpenAI	O OpenAI	92.7	-	-
#12	O GPT-5.5 Pro (batch)OpenAI	O OpenAI	92.7	-	367	-	New Hot
#13	O GPT-5.5OpenAI	O OpenAI	92.7	-	7		May change Falling
#14	O GPT-5.5 (batch)OpenAI	O OpenAI	92.7	-	365	-	New Hot
#15	G Gemini 3.1 Pro Preview Custom ToolsGoogle	G Google	92.2	-	8		May change Falling
#16	G Gemini 3.1 Pro PreviewGoogle	G Google	92.2	-	8		May change Falling
#17	G Gemini 3.1 Pro Preview (batch)Google	G Google	92.2	-	362	-	New Hot
#18	O GPT-5.4 ProOpenAI	O OpenAI	91.9	-	9		May change Falling
#19	O GPT-5.4 Pro (batch)OpenAI	O OpenAI	91.9	-	360	-	New Hot
#20	O GPT-5.4OpenAI	O OpenAI	91.9	-	10		May change Falling
#21	O GPT-5.4 (batch)OpenAI	O OpenAI	91.9	-	358	-	New Hot
#22	O GPT-5.3 ChatOpenAI	O OpenAI	90.5	-	10		May change Hot
#23	O GPT-5.3-CodexOpenAI	O OpenAI	90.5	-	-
#24	O GPT-5.2-CodexOpenAI	O OpenAI	90.5	-	12		May change Falling
#25	O GPT-5.2 ChatOpenAI	O OpenAI	90.5	-	40		May change Hot
#26	O GPT-5.2 ProOpenAI	O OpenAI	90.5	-	13		May change Falling
#27	O GPT-5.2 Pro (batch)OpenAI	O OpenAI	90.5	-	352	-	New Hot
#28	O GPT-5.2OpenAI	O OpenAI	90.5	-	14		May change Falling
#29	O GPT-5.2 (batch)OpenAI	O OpenAI	90.5	-	350	-	New Hot
#30	A Claude Opus 4.6Anthropic	A Anthropic	90.4	-	15		May change Falling
#31	A Claude Opus 4.6 (batch)Anthropic	A Anthropic	90.4	-	348	-	New Hot
#32	O GPT-5.6 Luna ProOpenAI	O OpenAI	89.0	-	16		May change Falling
#33	O GPT-5.6 Luna Pro (batch)OpenAI	O OpenAI	89.0	-	346	-	New Hot
#34	O GPT-5.6 LunaOpenAI	O OpenAI	89.0	-	17		May change Falling
#35	O GPT-5.6 Luna (batch)OpenAI	O OpenAI	89.0	-	344	-	New Hot
#36	O GPT-5.6 Terra ProOpenAI	O OpenAI	89.0	-	18		May change Falling
#37	O GPT-5.6 Terra Pro (batch)OpenAI	O OpenAI	89.0	-	342	-	New Hot
#38	O GPT-5.6 TerraOpenAI	O OpenAI	89.0	-	19		May change Falling
#39	O GPT-5.6 Terra (batch)OpenAI	O OpenAI	89.0	-	340	-	New Hot
#40	O GPT-5.6 Sol ProOpenAI	O OpenAI	89.0	-	20		May change Falling
#41	O GPT-5.6 Sol Pro (batch)OpenAI	O OpenAI	89.0	-	338	-	New Hot
#42	O GPT-5.6 SolOpenAI	O OpenAI	89.0	-	21		May change Falling
#43	O GPT-5.6 Sol (batch)OpenAI	O OpenAI	89.0	-	336	-	New Hot
#44	X Grok 4.5xAI	X xAI	88.8	-	335	-	New Hot
#45	X Grok 4.3xAI	X xAI	88.8	-	334	-	New Hot
#46	X Grok 4.20xAI	X xAI	88.8	-	333	-	New Hot
#47	O GPT-5.1-Codex-MaxOpenAI	O OpenAI	88.7	-	19		May change Falling
#48	O GPT-5.1OpenAI	O OpenAI	88.7	-	19		May change Falling
#49	O GPT-5.1-CodexOpenAI	O OpenAI	88.7	-	19		May change Falling
#50	O GPT-5 ProOpenAI	O OpenAI	88.7	-	26		May change Falling

Showing 50 of 378 models

Frequently Asked Questions

As of our latest rankings, Claude Fable 5 leads the best for coding category with a composite score of 97. Rankings are recalculated from benchmark, pricing, capability, and adoption signals as those sources refresh.

We use a composite scoring system that combines multiple signals: benchmark performance, Elo ratings, repository popularity, community sentiment, API latency, cost efficiency, adoption rates, and expert reviews. Each signal is normalized and weighted to produce a final score.

We currently track 378 AI models in the best for coding category. Our coverage is expanding as new models are released.

Ranking freshness depends on the underlying source. Pricing and provider catalog data refresh on a faster cadence, while benchmark and archived movement data update whenever new verified source data lands.

Yes! Click on any two models to see a detailed head-to-head comparison, including signal-by-signal breakdowns, pricing calculators, and personalized recommendations.