What is the most used AI model for coding?

Measured by real token burn from developers' machines on whoburnedmore, the top-ranked model on this page currently leads real coding usage. Unlike benchmark leaderboards, this ranking reflects what developers actually run in Claude Code, Codex, Cursor, Gemini CLI and other coding agents.

How is this different from OpenRouter's model rankings?

OpenRouter ranks models by API traffic routed through its gateway. This page measures usage that never touches an API gateway — subscription products like Claude Code Max plans, Cursor and Codex — read directly from the local usage logs those tools keep, via the open-source whoburnedmore CLI.

Where does this data come from?

From developers who run npx whoburnedmore and choose to be listed publicly. The CLI reads each coding agent's own local logs and submits daily aggregate token totals per tool and model — never prompts, code, or file names.

AI Model Usage Rankings — Real Coding Data, Not Benchmarks

#	Model	Devs	Tokens	Share	Input	Output	Cache	Est cost
01	gpt-5.6-solCodex CLI · OpenCode · Hermes +2	68	495.32B	26%	14.65B	1.36B	97%	$236,261
02	claude-opus-4-8Claude Code · OpenCode · Anthropic Api +3	130	346.33B	18%	5.57B	2.56B	98%	$369,549
03	gpt-5.5Codex CLI · OpenCode · Hermes +2	134	247.59B	13%	12.71B	891.6M	95%	$288,087
04	claude-opus-4-7Claude Code · OpenCode · Anthropic Api +1	74	91.15B	4.8%	138.3M	319.1M	100%	$75,576
05	claude-fable-5Claude Code · OpenCode	103	88.47B	4.7%	632.6M	664.1M	99%	$165,927
06	claude-opus-5Claude Code · Hermes · OpenCode	65	80.57B	4.3%	211.8M	356.6M	100%	$61,154
07	claude-sonnet-5Claude Code · Hermes	81	58.24B	3.1%	66.9M	233.2M	100%	$17,521
08	unreportedOpenCode · Factory Droid · Openclaw +9	86	55.68B	2.9%	8.35B	200.9M	84%	$35,019
09	defaultCursor · OpenCode	85	42.29B	2.2%	2.51B	259.8M	94%	$18,714
10	gpt-5.4Codex CLI · OpenCode · Hermes	97	34.24B	1.8%	3.73B	192.3M	89%	$25,449
11	claude-sonnet-4-6Claude Code · Cursor · Anthropic Api +4	114	29.93B	1.6%	552.9M	397.3M	98%	$21,330
12	gpt-5.6-terraCodex CLI · Hermes · Claude Code +1	69	23.26B	1.2%	915.9M	87.4M	96%	$7,756
13	gpt-5.3-codexCodex CLI · Cursor · OpenCode +2	86	21.27B	1.1%	1.67B	234.4M	92%	$9,602
14	gpt-5.6-lunaCodex CLI · OpenCode · Claude Code	45	16.44B	0.9%	567.7M	62.2M	97%	$1,703
15	claude-opus-4-6Claude Code · OpenCode · Amp +2	38	15.76B	0.8%	286.4M	163.9M	98%	$25,894
16	gpt-5Codex CLI · Cursor · OpenCode	60	14.00B	0.7%	560.9M	44.3M	96%	$1,540
17	claude-haiku-4-5-20251001Claude Code · Anthropic Api · Amp	129	11.22B	0.6%	65.1M	84.4M	99%	$2,510
18	claude-4.6-sonnet-medium-thinkingCursor	31	10.43B	0.6%	68.9M	78.4M	99%	$6,467
19	claude-4.6-opus-high-thinkingCursor	35	9.86B	0.5%	96.2M	54.4M	99%	$9,153
20	codex-auto-reviewCodex CLI	49	8.77B	0.5%	470.8M	22.1M	95%	$1,047
21	deepseek-v4-flash-freeOpenCode · Hermes · Claude Code	36	7.63B	0.4%	269.3M	33.9M	96%	$0.06
22	composer-2.5-fastCursor	47	7.47B	0.4%	475.9M	50.5M	94%	$5,614
23	gpt-5-codexCodex CLI · Cursor	25	7.43B	0.4%	315.3M	36.9M	96%	$2,166
24	gpt-5.5-mediumCursor	27	6.11B	0.3%	415.1M	13.2M	93%	$6,557
25	claude-4.5-opus-high-thinkingCursor	34	5.70B	0.3%	53.5M	38.7M	99%	$5,198
26	deepseek-v4-flashOpenCode · Claude Code · Kimi CLI +2	26	5.32B	0.3%	223.7M	18.5M	96%	$43.63
27	gpt-5.2-codexCodex CLI · OpenCode · Cursor	49	5.25B	0.3%	300.9M	38.1M	94%	$2,043
28	claude-4.5-opus-highCursor	8	5.22B	0.3%	55.5M	25.3M	99%	$4,607
29	gpt-realtime-1.5Openai Api	1	4.98B	0.3%	2.81B	7.2M	43%	$0
30	gpt-5.5-highCursor	14	4.76B	0.3%	298.0M	19.5M	94%	$4,235
31	composer-2-fastCursor	40	4.64B	0.2%	355.2M	30.3M	92%	$1,953
32	claude-opus-4-8-thinking-highCursor	25	4.45B	0.2%	19.3M	23.5M	100%	$4,174
33	claude-4-sonnet-thinkingCursor	27	4.44B	0.2%	42.2M	39.9M	99%	$3,266
34	cursor-grok-4.5-high-fastCursor	18	4.08B	0.2%	364.2M	32.1M	91%	$5,616
35	rollupClaude Code	4	4.00B	0.2%	4.00B	0	0.0%	$3,429
36	deepseek-v4-proClaude Code · OpenCode · Hermes +2	23	3.56B	0.2%	53.3M	7.2M	98%	$85.83
37	claude-opus-4-7-thinking-xhighCursor	21	3.39B	0.2%	12.6M	15.1M	100%	$3,262
38	claude-4-sonnetCursor	29	3.25B	0.2%	30.8M	18.8M	99%	$1,932
39	claude-4.6-sonnet-high-thinkingCursor	6	3.23B	0.2%	31.1M	32.3M	99%	$1,959
40	composer-2.5Cursor	20	2.92B	0.2%	188.2M	23.3M	94%	$701
41	claude-4.5-sonnetCursor	21	2.86B	0.2%	22.7M	21.5M	99%	$2,084
42	claude-opus-4-5-20251101OpenCode · Claude Code · Amp +1	15	2.82B	0.1%	220.5M	12.0M	92%	$3,548
43	claude-4.5-sonnet-thinkingCursor	36	2.64B	0.1%	55.4M	26.7M	98%	$2,235
44	premium (codex 5.3)Cursor	11	2.64B	0.1%	262.2M	9.0M	90%	$931
45	claude-opus-4-7-thinking-highCursor	19	2.61B	0.1%	11.0M	10.2M	100%	$2,126
46	rooRoo Code	3	2.38B	0.1%	1.47B	13.6M	35%	$1,157
47	gpt-5.4-miniCodex CLI · Hermes · Cursor +1	54	2.28B	0.1%	215.7M	13.9M	90%	$703
48	gpt-5.5-high-fastCursor	3	2.09B	0.1%	179.6M	10.3M	91%	$4,492
49	agent_reviewCursor	36	2.08B	0.1%	22.1M	10.4M	99%	$499
50	premiumCursor	4	2.04B	0.1%	182.7M	5.7M	91%	$682
51	gpt-5.3-codex-highCursor · Codex CLI	14	1.92B	0.1%	443.2M	236.5M	72%	$3,276
52	glm-5.2OpenCode · Hermes · Codex CLI +2	20	1.85B	<0.1%	340.6M	10.8M	81%	$851
53	k3Kimi CLI · Claude Code · OpenCode	4	1.83B	<0.1%	60.7M	9.6M	97%	$840
54	minimax-m3Codex CLI · OpenCode · Hermes +1	12	1.63B	<0.1%	105.6M	6.3M	93%	$175
55	claude-4.6-opus-highCursor	6	1.46B	<0.1%	17.3M	7.9M	99%	$1,365
56	claude-opus-4-7-thinking-mediumCursor	7	1.37B	<0.1%	7.6M	6.1M	99%	$1,209
57	claude-fable-5-thinking-maxCursor	3	1.35B	<0.1%	741.2K	7.2M	100%	$2,748
58	claude-fable-5-thinking-highCursor	15	1.33B	<0.1%	3.4M	7.1M	100%	$2,847
59	gpt-5.2Codex CLI · Cursor · OpenCode	29	1.29B	<0.1%	95.7M	10.9M	93%	$519
60	gpt-5.4-highCursor	6	1.25B	<0.1%	139.8M	3.7M	89%	$1,131

AI MODEL USAGE RANKINGS — real coding burn, all time

Models by real token burn

Daily burn × model

AI MODEL USAGE RANKINGS — real coding burn, all time

▮Models by real token burn

▮Daily burn × model

Models by real token burn

Daily burn × model