The leaders, at a glance

Agentic AI model ranking

Compare AI models and effort settings for agentic work using sourced task evaluations and visible evidence coverage.

See the full ranking ↓How configurations are ranked

Agentic leaders

Experimental comparison score · 0–100

For the selected profile and sources. Scores are experimental; small gaps may not be meaningful. How scoring works · Effort evidence collected 2026-09-12

451unique modelsEach model counted once across effort levels 492model configurationsDocumented model + effort settings, including unranked

109aggregated sourcesOriginal reports and benchmark leaderboards

Explore all 109 sources Independent evaluations and lab reports

Distinct source pages with published results in the catalog. Multiple results and page sections count as one source.

Every ranked row is a measured configuration. The default view shows the highest-scoring supported configuration per model. In Value, the budget is applied first. Select All efforts for the full comparison. How effort ranking works.

Customize ranking

Full ranking

127 entries

Custom capability ranking of measured model configurations. Scores are experimental; missing results remain unknown.
Rank	Model	Score	Capability coverage	analyst-agent	briefcase	aime	apex-agents	aa-lcr	automationbench	critpt	enterprise-ops	gdp-pdf	gdpval	gmmlu	gpqa	harvey	hle	ifbench	itbench	livecodebench	math500	mmlu-pro	mmmu-pro	omniscience	scicode	tau-bench	terminal-bench	arc-agi	deepswe	swe-bench-pro	arc-agi-3	frontiercode	enigma-eval	swe-atlas-qna	swe-atlas-refactoring	swe-atlas-test-writing	livebench-math	livebench-coding	livebench-language	livebench-data	multi-swe-bench	livebench-reasoning	livebench-instructions	swe-rebench	epoch-game-puzzles	frontiermath	simpleqa	vals-finance-agent	vals-code-migration	vals-legal-research	vals-excel-modeling	vibe-code	chartography	osworld	charxiv	swe-bench-multilingual	deepsearchqa	officeqa	mcp-atlas	toolathlon	mrcr	browsecomp
1	Claude Fable 5.1 · MaxAnthropic Measured effort: Max	79.6	Eligible · 100%63 published · 12 families · 1 selected capabilities	96.928/28 peersIndependent evidence	100.087/87 peersIndependent evidence	98.448/48 peersIndependent evidence	—	99.9198/198 peersIndependent evidence	99.589/89 peersIndependent evidence	98.6200/200 peersIndependent evidence	—	97.188/88 peersIndependent evidence	100.0123/123 peersIndependent evidence	—	98.4212/212 peersIndependent evidence	66.550/50 peersIndependent evidence	100.0208/208 peersIndependent evidence	—	—	—	—	100.031/31 peersIndependent evidence	—	100.0198/198 peersIndependent evidence	100.091/91 peersIndependent evidence	94.2112/112 peersIndependent evidence	100.0125/125 peersIndependent evidence	97.738/38 peersIndependent evidence	—	—	—	87.47/7 peersIndependent evidence	—	—	—	—	91.626/26 peersIndependent evidence	91.126/26 peersIndependent evidence	79.526/26 peersIndependent evidence	64.226/26 peersIndependent evidence	87.826/26 peersIndependent evidence	77.226/26 peersIndependent evidence	66.426/26 peersIndependent evidence	—	93.960/60 peersIndependent evidence	95.739/39 peersIndependent evidence	97.025/25 peersIndependent evidence	87.030/30 peersIndependent evidence	90.733/33 peersIndependent evidence	98.332/32 peersIndependent evidence	100.032/32 peersIndependent evidence	96.732/32 peersIndependent evidence	100.02/2 peersPublisher reports	100.02/2 peersPublisher reports	—	—	—	—	—	—	—	—
2	Muse Spark 1.3 · XHighMeta Measured effort: XHigh	77.2	Eligible · 100%49 published · 11 families · 1 selected capabilities	—	93.887/87 peersIndependent evidence	—	—	98.3197/197 peersIndependent evidence	98.389/89 peersIndependent evidence	97.0200/200 peersIndependent evidence	—	96.088/88 peersIndependent evidence	98.8123/123 peersIndependent evidence	—	99.1212/212 peersIndependent evidence	98.250/50 peersIndependent evidence	98.1208/208 peersIndependent evidence	—	—	—	—	—	93.7114/114 peersIndependent evidence	90.0198/198 peersIndependent evidence	98.891/91 peersIndependent evidence	95.3112/112 peersIndependent evidence	76.7126/126 peersIndependent evidence	—	—	—	—	—	—	—	—	—	82.626/26 peersIndependent evidence	76.526/26 peersIndependent evidence	56.626/26 peersIndependent evidence	75.926/26 peersIndependent evidence	89.726/26 peersIndependent evidence	74.726/26 peersIndependent evidence	90.526/26 peersIndependent evidence	—	—	—	—	92.730/30 peersIndependent evidence	—	70.632/32 peersIndependent evidence	57.632/32 peersIndependent evidence	93.232/32 peersIndependent evidence	—	—	—	—	100.01/1 peersPublisher reports	—	—	—	100.01/1 peersPublisher reports	—
3	Kimi K3 · MaxMoonshot Measured effort: Max	76.2	Eligible · 100%49 published · 16 families · 1 selected capabilities	59.028/28 peersIndependent evidence	93.987/87 peersIndependent evidence	86.048/48 peersIndependent evidence	97.226/26 peersIndependent evidence	93.2198/198 peersIndependent evidence	98.289/89 peersIndependent evidence	96.1200/200 peersIndependent evidence	80.541/41 peersIndependent evidence	91.888/88 peersIndependent evidence	92.7123/123 peersIndependent evidence	—	82.3213/213 peersIndependent evidence	97.037/37 peersIndependent evidence	52.8208/208 peersIndependent evidence	—	95.228/28 peersIndependent evidence	—	—	59.231/31 peersIndependent evidence	74.4115/115 peersIndependent evidence	92.5198/198 peersIndependent evidence	98.191/91 peersIndependent evidence	94.5112/112 peersIndependent evidence	89.0125/125 peersIndependent evidence	66.938/38 peersIndependent evidence	79.025/25 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	74.860/60 peersIndependent evidence	70.539/39 peersIndependent evidence	54.525/25 peersIndependent evidence	—	29.733/33 peersIndependent evidence	—	80.732/32 peersIndependent evidence	—	—	83.33/3 peersPublisher reports	100.03/3 peersPublisher reports	—	100.01/1 peersPublisher reports	75.04/4 peersPublisher reports	100.04/4 peersPublisher reports	100.04/4 peersPublisher reports	—	—
4	GPT-6 Astra · XHighOpenAI Measured effort: XHigh	73.7	Eligible · 100%26 published · 5 families · 1 selected capabilities	—	95.787/87 peersIndependent evidence	—	—	93.7197/197 peersIndependent evidence	100.089/89 peersIndependent evidence	99.9200/200 peersIndependent evidence	—	100.088/88 peersIndependent evidence	92.7123/123 peersIndependent evidence	—	100.0212/212 peersIndependent evidence	—	99.6208/208 peersIndependent evidence	—	—	—	—	—	100.0114/114 peersIndependent evidence	99.5198/198 peersIndependent evidence	91.191/91 peersIndependent evidence	88.4112/112 peersIndependent evidence	99.8125/125 peersIndependent evidence	100.038/38 peersIndependent evidence	100.025/25 peersIndependent evidence	—	100.06/6 peersIndependent evidence	100.05/5 peersIndependent evidence	—	—	100.02/2 peersIndependent evidence	100.02/2 peersIndependent evidence	—	—	—	—	—	—	—	—	—	100.029/29 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
5	Grok 4.6 · XHighxAI Measured effort: XHigh	68.5	Eligible · 100%25 published · 5 families · 1 selected capabilities	—	95.587/87 peersIndependent evidence	93.548/48 peersIndependent evidence	—	96.3197/197 peersIndependent evidence	99.689/89 peersIndependent evidence	94.7200/200 peersIndependent evidence	—	83.988/88 peersIndependent evidence	98.6123/123 peersIndependent evidence	—	95.1213/213 peersIndependent evidence	—	97.1208/208 peersIndependent evidence	—	—	—	—	—	—	91.8198/198 peersIndependent evidence	83.491/91 peersIndependent evidence	88.0112/112 peersIndependent evidence	92.0125/125 peersIndependent evidence	75.338/38 peersIndependent evidence	70.825/25 peersIndependent evidence	—	60.36/6 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	78.060/60 peersIndependent evidence	62.339/39 peersIndependent evidence	55.925/25 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
6	Claude Opus 5 · XHighAnthropic Measured effort: XHigh	68.2	Eligible · 100%21 published · 5 families · 1 selected capabilities	—	99.487/87 peersIndependent evidence	—	—	96.2198/198 peersIndependent evidence	84.989/89 peersIndependent evidence	97.8200/200 peersIndependent evidence	—	91.088/88 peersIndependent evidence	99.6123/123 peersIndependent evidence	—	99.0212/212 peersIndependent evidence	—	99.5208/208 peersIndependent evidence	—	—	—	—	—	97.8114/114 peersIndependent evidence	99.0198/198 peersIndependent evidence	91.091/91 peersIndependent evidence	87.5112/112 peersIndependent evidence	97.7125/125 peersIndependent evidence	—	96.625/25 peersIndependent evidence	—	—	67.67/7 peersIndependent evidence	—	100.02/2 peersIndependent evidence	—	100.02/2 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
7	GPT-5.6 Sol · MaxOpenAI Measured effort: Max	67.2	Eligible · 100%81 published · 17 families · 1 selected capabilities	80.428/28 peersIndependent evidence	92.387/87 peersIndependent evidence	98.448/48 peersIndependent evidence	—	92.8198/198 peersIndependent evidence	81.189/89 peersIndependent evidence	100.0200/200 peersIndependent evidence	65.941/41 peersIndependent evidence	98.488/88 peersIndependent evidence	93.3123/123 peersIndependent evidence	—	97.3213/213 peersIndependent evidence	43.550/50 peersIndependent evidence	86.3208/208 peersIndependent evidence	87.7178/178 peersIndependent evidence	100.028/28 peersIndependent evidence	—	—	61.231/31 peersIndependent evidence	99.3115/115 peersIndependent evidence	98.6198/198 peersIndependent evidence	93.891/91 peersIndependent evidence	84.2202/202 peersIndependent evidence	74.6201/201 peersIndependent evidence	98.938/38 peersIndependent evidence	92.425/25 peersIndependent evidence	—	70.96/6 peersIndependent evidence	85.75/5 peersIndependent evidence	—	—	—	—	63.426/26 peersIndependent evidence	90.326/26 peersIndependent evidence	79.726/26 peersIndependent evidence	72.026/26 peersIndependent evidence	63.426/26 peersIndependent evidence	82.326/26 peersIndependent evidence	61.126/26 peersIndependent evidence	—	97.860/60 peersIndependent evidence	92.339/39 peersIndependent evidence	66.76/6 peersIndependent evidence	49.930/30 peersIndependent evidence	87.833/33 peersIndependent evidence	85.932/32 peersIndependent evidence	90.632/32 peersIndependent evidence	78.132/32 peersIndependent evidence	—	68.33/3 peersPublisher reports	53.33/3 peersPublisher reports	—	—	53.64/4 peersPublisher reports	65.84/4 peersPublisher reports	53.64/4 peersPublisher reports	—	—
8	Claude Fable 5 · MaxAnthropic Measured effort: Max	65.9	Eligible · 100%67 published · 13 families · 1 selected capabilities	83.828/28 peersIndependent evidence	95.687/87 peersIndependent evidence	100.01/1 peersIndependent evidence	—	97.5198/198 peersIndependent evidence	84.489/89 peersIndependent evidence	97.9200/200 peersIndependent evidence	100.041/41 peersIndependent evidence	95.288/88 peersIndependent evidence	95.3123/123 peersIndependent evidence	—	68.3213/213 peersIndependent evidence	94.537/37 peersIndependent evidence	99.7208/208 peersIndependent evidence	73.6178/178 peersIndependent evidence	—	—	—	93.231/31 peersIndependent evidence	—	99.8198/198 peersIndependent evidence	99.891/91 peersIndependent evidence	92.0202/202 peersIndependent evidence	39.3201/201 peersIndependent evidence	96.138/38 peersIndependent evidence	84.025/25 peersIndependent evidence	—	—	97.97/7 peersIndependent evidence	—	—	—	—	78.126/26 peersIndependent evidence	82.926/26 peersIndependent evidence	80.226/26 peersIndependent evidence	64.726/26 peersIndependent evidence	83.926/26 peersIndependent evidence	63.626/26 peersIndependent evidence	74.026/26 peersIndependent evidence	—	93.940/40 peersIndependent evidence	93.739/39 peersIndependent evidence	—	73.230/30 peersIndependent evidence	94.033/33 peersIndependent evidence	93.432/32 peersIndependent evidence	96.732/32 peersIndependent evidence	100.032/32 peersIndependent evidence	49.62/2 peersPublisher reports	0.02/2 peersPublisher reports	—	—	—	—	—	—	—	—
9	Gemini 3.7 Flash · MediumGoogle Measured effort: Medium	64.0	Eligible · 100%16 published · 3 families · 1 selected capabilities	—	—	—	—	98.3197/197 peersIndependent evidence	—	87.8200/200 peersIndependent evidence	98.941/41 peersIndependent evidence	—	87.8123/123 peersIndependent evidence	—	94.1212/212 peersIndependent evidence	—	91.9208/208 peersIndependent evidence	—	—	—	—	—	99.1114/114 peersIndependent evidence	97.2198/198 peersIndependent evidence	99.391/91 peersIndependent evidence	81.4112/112 peersIndependent evidence	86.5125/125 peersIndependent evidence	74.538/38 peersIndependent evidence	66.525/25 peersIndependent evidence	—	—	100.06/6 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
10	Qwen3.8-2.4T-A95B · ReasoningQwen Measured effort: Reasoning	63.8	Eligible · 100%16 published · 6 families · 1 selected capabilities	—	89.687/87 peersIndependent evidence	—	—	58.6198/198 peersIndependent evidence	92.789/89 peersIndependent evidence	95.4200/200 peersIndependent evidence	95.041/41 peersIndependent evidence	80.588/88 peersIndependent evidence	94.0123/123 peersIndependent evidence	—	98.3212/212 peersIndependent evidence	—	95.6208/208 peersIndependent evidence	—	—	—	—	—	—	81.6198/198 peersIndependent evidence	86.891/91 peersIndependent evidence	96.8112/112 peersIndependent evidence	84.8125/125 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
11	Qwen3.8-Flash-Next · ReasoningQwen Measured effort: Reasoning	63.2	Eligible · 100%15 published · 5 families · 1 selected capabilities	—	98.187/87 peersIndependent evidence	—	—	93.5197/197 peersIndependent evidence	88.889/89 peersIndependent evidence	90.2200/200 peersIndependent evidence	—	80.588/88 peersIndependent evidence	96.8123/123 peersIndependent evidence	—	95.2212/212 peersIndependent evidence	—	91.6208/208 peersIndependent evidence	—	—	—	—	—	91.2114/114 peersIndependent evidence	67.3198/198 peersIndependent evidence	75.091/91 peersIndependent evidence	90.3112/112 peersIndependent evidence	93.7125/125 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
12	Qwen 3.8-Max · ReasoningQwen Measured effort: Reasoning	61.9	Eligible · 100%16 published · 5 families · 1 selected capabilities	—	87.087/87 peersIndependent evidence	—	—	83.7198/198 peersIndependent evidence	81.989/89 peersIndependent evidence	95.4200/200 peersIndependent evidence	—	90.088/88 peersIndependent evidence	95.2123/123 peersIndependent evidence	—	96.2212/212 peersIndependent evidence	—	96.7208/208 peersIndependent evidence	—	—	—	—	—	95.7114/114 peersIndependent evidence	82.1198/198 peersIndependent evidence	84.491/91 peersIndependent evidence	100.0112/112 peersIndependent evidence	87.6125/125 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
13	Gemini 3.8 Flash · MediumGoogle Measured effort: Medium	61.3	Eligible · 100%18 published · 5 families · 1 selected capabilities	—	76.787/87 peersIndependent evidence	—	—	99.7197/197 peersIndependent evidence	97.589/89 peersIndependent evidence	90.8200/200 peersIndependent evidence	—	93.388/88 peersIndependent evidence	89.0123/123 peersIndependent evidence	—	98.2212/212 peersIndependent evidence	—	94.7208/208 peersIndependent evidence	—	—	—	—	—	98.3114/114 peersIndependent evidence	96.6198/198 peersIndependent evidence	90.091/91 peersIndependent evidence	91.1112/112 peersIndependent evidence	89.8125/125 peersIndependent evidence	—	90.025/25 peersIndependent evidence	—	—	85.66/6 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
14	GLM-5.3 · MaxZ.ai Measured effort: Max	60.1	Eligible · 100%36 published · 10 families · 1 selected capabilities	—	94.887/87 peersIndependent evidence	61.948/48 peersIndependent evidence	—	90.8198/198 peersIndependent evidence	98.289/89 peersIndependent evidence	93.5200/200 peersIndependent evidence	39.741/41 peersIndependent evidence	62.788/88 peersIndependent evidence	98.9123/123 peersIndependent evidence	—	80.5213/213 peersIndependent evidence	74.132/32 peersIndependent evidence	92.2208/208 peersIndependent evidence	—	—	—	—	30.131/31 peersIndependent evidence	—	80.4198/198 peersIndependent evidence	97.491/91 peersIndependent evidence	98.4112/112 peersIndependent evidence	50.2126/126 peersIndependent evidence	—	85.325/25 peersIndependent evidence	—	—	80.66/6 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	63.060/60 peersIndependent evidence	56.939/39 peersIndependent evidence	29.225/25 peersIndependent evidence	66.630/30 peersIndependent evidence	68.433/33 peersIndependent evidence	90.032/32 peersIndependent evidence	32.732/32 peersIndependent evidence	68.332/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—
15	GPT-5.6 Terra · MaxOpenAI Measured effort: Max	59.3	Eligible · 100%67 published · 13 families · 1 selected capabilities	—	85.287/87 peersIndependent evidence	96.148/48 peersIndependent evidence	92.226/26 peersIndependent evidence	93.5198/198 peersIndependent evidence	94.889/89 peersIndependent evidence	98.7200/200 peersIndependent evidence	39.041/41 peersIndependent evidence	94.888/88 peersIndependent evidence	90.1123/123 peersIndependent evidence	—	92.4213/213 peersIndependent evidence	35.750/50 peersIndependent evidence	95.3208/208 peersIndependent evidence	84.2178/178 peersIndependent evidence	96.428/28 peersIndependent evidence	—	—	—	92.6114/114 peersIndependent evidence	93.0198/198 peersIndependent evidence	87.791/91 peersIndependent evidence	84.1202/202 peersIndependent evidence	42.6201/201 peersIndependent evidence	85.938/38 peersIndependent evidence	81.525/25 peersIndependent evidence	—	45.06/6 peersIndependent evidence	64.25/5 peersIndependent evidence	—	—	—	—	56.326/26 peersIndependent evidence	47.526/26 peersIndependent evidence	49.126/26 peersIndependent evidence	69.626/26 peersIndependent evidence	59.626/26 peersIndependent evidence	77.426/26 peersIndependent evidence	26.926/26 peersIndependent evidence	—	91.360/60 peersIndependent evidence	81.539/39 peersIndependent evidence	33.36/6 peersIndependent evidence	59.930/30 peersIndependent evidence	84.933/33 peersIndependent evidence	68.532/32 peersIndependent evidence	66.132/32 peersIndependent evidence	56.132/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—
16	Claude Opus 4.8 · MaxAnthropic Measured effort: Max	58.9	Eligible · 100%78 published · 17 families · 1 selected capabilities	74.628/28 peersIndependent evidence	84.487/87 peersIndependent evidence	0.01/1 peersIndependent evidence	—	85.4198/198 peersIndependent evidence	58.789/89 peersIndependent evidence	95.3200/200 peersIndependent evidence	70.041/41 peersIndependent evidence	92.788/88 peersIndependent evidence	89.0123/123 peersIndependent evidence	—	25.5212/212 peersIndependent evidence	82.750/50 peersIndependent evidence	87.1208/208 peersIndependent evidence	71.8178/178 peersIndependent evidence	—	—	—	77.431/31 peersIndependent evidence	0.03/3 peersPublisher reports	92.3198/198 peersIndependent evidence	84.691/91 peersIndependent evidence	85.5202/202 peersIndependent evidence	31.1201/201 peersIndependent evidence	—	25.84/4 peersPublisher reports	—	—	66.67/7 peersIndependent evidence	—	—	—	—	74.826/26 peersIndependent evidence	71.926/26 peersIndependent evidence	40.326/26 peersIndependent evidence	32.026/26 peersIndependent evidence	46.226/26 peersIndependent evidence	62.826/26 peersIndependent evidence	61.726/26 peersIndependent evidence	—	37.240/40 peersIndependent evidence	77.739/39 peersIndependent evidence	68.025/25 peersIndependent evidence	57.230/30 peersIndependent evidence	82.333/33 peersIndependent evidence	76.232/32 peersIndependent evidence	79.232/32 peersIndependent evidence	87.132/32 peersIndependent evidence	—	52.03/3 peersPublisher reports	34.63/3 peersPublisher reports	—	0.01/1 peersPublisher reports	100.04/4 peersPublisher reports	64.44/4 peersPublisher reports	77.14/4 peersPublisher reports	—	—
17	GLM-5.3-Flash · MaxZ.ai Measured effort: Max	58.5	Eligible · 100%17 published · 4 families · 1 selected capabilities	—	—	72.748/48 peersIndependent evidence	—	—	—	—	—	—	—	—	77.447/47 peersIndependent evidence	57.632/32 peersIndependent evidence	—	—	—	—	—	22.731/31 peersIndependent evidence	—	—	—	—	34.729/29 peersIndependent evidence	—	63.825/25 peersIndependent evidence	—	—	41.26/6 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	31.060/60 peersIndependent evidence	36.539/39 peersIndependent evidence	—	81.530/30 peersIndependent evidence	30.333/33 peersIndependent evidence	83.932/32 peersIndependent evidence	33.732/32 peersIndependent evidence	16.532/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—
18	Claude Sonnet 5 · MaxAnthropic Measured effort: Max	57.1	Eligible · 100%38 published · 11 families · 1 selected capabilities	79.828/28 peersIndependent evidence	86.287/87 peersIndependent evidence	49.448/48 peersIndependent evidence	—	95.8198/198 peersIndependent evidence	75.989/89 peersIndependent evidence	92.8200/200 peersIndependent evidence	78.541/41 peersIndependent evidence	73.788/88 peersIndependent evidence	91.4123/123 peersIndependent evidence	—	49.2213/213 peersIndependent evidence	55.150/50 peersIndependent evidence	93.7208/208 peersIndependent evidence	—	—	—	—	43.831/31 peersIndependent evidence	80.3114/114 peersIndependent evidence	89.2198/198 peersIndependent evidence	87.791/91 peersIndependent evidence	85.6112/112 peersIndependent evidence	12.3126/126 peersIndependent evidence	—	37.325/25 peersIndependent evidence	—	—	50.87/7 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	67.760/60 peersIndependent evidence	51.639/39 peersIndependent evidence	24.725/25 peersIndependent evidence	53.630/30 peersIndependent evidence	71.633/33 peersIndependent evidence	71.232/32 peersIndependent evidence	70.032/32 peersIndependent evidence	80.432/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—
19	Qwen3.8-27B · XHighQwen Measured effort: XHigh	56.8	Eligible · 100%18 published · 6 families · 1 selected capabilities	—	87.187/87 peersIndependent evidence	—	—	90.7198/198 peersIndependent evidence	79.689/89 peersIndependent evidence	85.3200/200 peersIndependent evidence	83.441/41 peersIndependent evidence	80.688/88 peersIndependent evidence	90.7123/123 peersIndependent evidence	—	93.6212/212 peersIndependent evidence	—	90.2208/208 peersIndependent evidence	—	—	—	—	—	85.1114/114 peersIndependent evidence	29.2198/198 peersIndependent evidence	59.391/91 peersIndependent evidence	94.6112/112 peersIndependent evidence	79.9125/125 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
20	GPT-5.5 · XHighOpenAI Measured effort: XHigh	56.2	Eligible · 100%83 published · 18 families · 1 selected capabilities	87.228/28 peersIndependent evidence	74.687/87 peersIndependent evidence	—	88.526/26 peersIndependent evidence	84.9198/198 peersIndependent evidence	36.089/89 peersIndependent evidence	97.3200/200 peersIndependent evidence	81.641/41 peersIndependent evidence	91.188/88 peersIndependent evidence	85.1123/123 peersIndependent evidence	—	68.4212/212 peersIndependent evidence	47.250/50 peersIndependent evidence	15.1208/208 peersIndependent evidence	93.6178/178 peersIndependent evidence	85.728/28 peersIndependent evidence	—	—	48.431/31 peersIndependent evidence	42.5115/115 peersIndependent evidence	98.2198/198 peersIndependent evidence	89.091/91 peersIndependent evidence	87.6202/202 peersIndependent evidence	67.9201/201 peersIndependent evidence	90.738/38 peersIndependent evidence	66.025/25 peersIndependent evidence	—	—	68.05/5 peersIndependent evidence	—	50.02/2 peersIndependent evidence	50.02/2 peersIndependent evidence	0.02/2 peersIndependent evidence	56.326/26 peersIndependent evidence	80.826/26 peersIndependent evidence	79.526/26 peersIndependent evidence	79.526/26 peersIndependent evidence	55.126/26 peersIndependent evidence	65.926/26 peersIndependent evidence	55.826/26 peersIndependent evidence	100.06/6 peersIndependent evidence	98.140/40 peersIndependent evidence	82.039/39 peersIndependent evidence	74.125/25 peersIndependent evidence	40.030/30 peersIndependent evidence	78.833/33 peersIndependent evidence	59.432/32 peersIndependent evidence	59.432/32 peersIndependent evidence	43.832/32 peersIndependent evidence	—	0.03/3 peersPublisher reports	16.83/3 peersPublisher reports	—	—	28.54/4 peersPublisher reports	28.54/4 peersPublisher reports	28.54/4 peersPublisher reports	—	—
21	Muse Spark 1.2 · XHighMeta Measured effort: XHigh	55.8	Eligible · 100%54 published · 12 families · 1 selected capabilities	—	86.287/87 peersIndependent evidence	—	—	85.5198/198 peersIndependent evidence	12.989/89 peersIndependent evidence	92.1200/200 peersIndependent evidence	85.541/41 peersIndependent evidence	82.888/88 peersIndependent evidence	91.4123/123 peersIndependent evidence	—	88.3212/212 peersIndependent evidence	100.032/32 peersIndependent evidence	95.7208/208 peersIndependent evidence	—	—	—	—	55.231/31 peersIndependent evidence	—	90.1198/198 peersIndependent evidence	93.191/91 peersIndependent evidence	83.9112/112 peersIndependent evidence	62.9126/126 peersIndependent evidence	—	51.825/25 peersIndependent evidence	—	—	—	—	—	—	—	43.926/26 peersIndependent evidence	43.626/26 peersIndependent evidence	49.826/26 peersIndependent evidence	56.326/26 peersIndependent evidence	69.026/26 peersIndependent evidence	55.526/26 peersIndependent evidence	68.726/26 peersIndependent evidence	—	—	—	82.525/25 peersIndependent evidence	97.230/30 peersIndependent evidence	33.733/33 peersIndependent evidence	77.232/32 peersIndependent evidence	45.032/32 peersIndependent evidence	75.732/32 peersIndependent evidence	—	—	—	—	0.01/1 peersPublisher reports	—	—	—	0.01/1 peersPublisher reports	—
22	GPT-5.6 Luna · MaxOpenAI Measured effort: Max	55.1	Eligible · 100%65 published · 13 families · 1 selected capabilities	—	85.487/87 peersIndependent evidence	88.448/48 peersIndependent evidence	85.726/26 peersIndependent evidence	89.1198/198 peersIndependent evidence	82.789/89 peersIndependent evidence	95.4200/200 peersIndependent evidence	54.941/41 peersIndependent evidence	94.588/88 peersIndependent evidence	90.5123/123 peersIndependent evidence	—	85.2213/213 peersIndependent evidence	41.550/50 peersIndependent evidence	91.1208/208 peersIndependent evidence	—	72.528/28 peersIndependent evidence	—	—	16.231/31 peersIndependent evidence	85.9114/114 peersIndependent evidence	89.9198/198 peersIndependent evidence	83.791/91 peersIndependent evidence	80.0112/112 peersIndependent evidence	13.1126/126 peersIndependent evidence	62.538/38 peersIndependent evidence	71.125/25 peersIndependent evidence	—	10.46/6 peersIndependent evidence	55.55/5 peersIndependent evidence	—	—	—	—	24.526/26 peersIndependent evidence	73.326/26 peersIndependent evidence	14.926/26 peersIndependent evidence	61.926/26 peersIndependent evidence	29.626/26 peersIndependent evidence	32.326/26 peersIndependent evidence	16.526/26 peersIndependent evidence	—	71.360/60 peersIndependent evidence	78.339/39 peersIndependent evidence	16.76/6 peersIndependent evidence	63.230/30 peersIndependent evidence	72.933/33 peersIndependent evidence	37.632/32 peersIndependent evidence	74.932/32 peersIndependent evidence	65.532/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—
23	Claude Opus 4.7 · MaxAnthropic Measured effort: Max	53.1	Eligible · 100%32 published · 8 families · 1 selected capabilities	75.528/28 peersIndependent evidence	80.887/87 peersIndependent evidence	59.348/48 peersIndependent evidence	—	87.5197/197 peersIndependent evidence	—	88.3200/200 peersIndependent evidence	—	—	83.8123/123 peersIndependent evidence	—	66.7213/213 peersIndependent evidence	57.732/32 peersIndependent evidence	93.5208/208 peersIndependent evidence	68.8178/178 peersIndependent evidence	93.428/28 peersIndependent evidence	—	—	84.231/31 peersIndependent evidence	86.3114/114 peersIndependent evidence	93.8198/198 peersIndependent evidence	—	82.3202/202 peersIndependent evidence	36.5201/201 peersIndependent evidence	—	—	—	—	39.57/7 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	48.660/60 peersIndependent evidence	66.639/39 peersIndependent evidence	—	—	71.433/33 peersIndependent evidence	52.432/32 peersIndependent evidence	64.232/32 peersIndependent evidence	54.132/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—
24	GPT-5.4 · XHighOpenAI Measured effort: XHigh	52.1	Eligible · 100%50 published · 5 families · 1 selected capabilities	—	—	—	84.726/26 peersIndependent evidence	96.6197/197 peersIndependent evidence	—	96.3200/200 peersIndependent evidence	—	—	81.0123/123 peersIndependent evidence	—	93.0212/212 peersIndependent evidence	9.832/32 peersIndependent evidence	96.2208/208 peersIndependent evidence	90.1178/178 peersIndependent evidence	—	—	—	47.831/31 peersIndependent evidence	83.6114/114 peersIndependent evidence	94.7198/198 peersIndependent evidence	—	84.9202/202 peersIndependent evidence	16.9201/201 peersIndependent evidence	82.838/38 peersIndependent evidence	32.025/25 peersIndependent evidence	100.02/2 peersIndependent evidence	—	—	—	0.02/2 peersIndependent evidence	0.02/2 peersIndependent evidence	25.03/3 peersIndependent evidence	49.026/26 peersIndependent evidence	42.526/26 peersIndependent evidence	62.026/26 peersIndependent evidence	73.926/26 peersIndependent evidence	55.626/26 peersIndependent evidence	55.926/26 peersIndependent evidence	53.026/26 peersIndependent evidence	—	—	—	—	—	49.333/33 peersIndependent evidence	—	—	45.932/32 peersIndependent evidence	—	—	100.02/2 peersPublisher reports	—	—	—	—	—	—	—
25	DeepSeek V4-Pro 0813 · MaxDeepSeek Measured effort: Max	50.9	Eligible · 100%33 published · 10 families · 1 selected capabilities	—	82.387/87 peersIndependent evidence	89.148/48 peersIndependent evidence	—	82.9198/198 peersIndependent evidence	90.289/89 peersIndependent evidence	93.1200/200 peersIndependent evidence	94.641/41 peersIndependent evidence	66.588/88 peersIndependent evidence	90.4123/123 peersIndependent evidence	—	86.8213/213 peersIndependent evidence	70.932/32 peersIndependent evidence	91.4208/208 peersIndependent evidence	—	—	—	—	38.131/31 peersIndependent evidence	—	93.7198/198 peersIndependent evidence	74.291/91 peersIndependent evidence	88.8112/112 peersIndependent evidence	50.2126/126 peersIndependent evidence	73.138/38 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	91.160/60 peersIndependent evidence	48.839/39 peersIndependent evidence	65.625/25 peersIndependent evidence	36.430/30 peersIndependent evidence	66.833/33 peersIndependent evidence	66.432/32 peersIndependent evidence	27.432/32 peersIndependent evidence	85.332/32 peersIndependent evidence	—	—	—	—	—	—	—	—	—	—

About this ranking

This profile gives agentic work all the weight. Published agent-task evaluations contribute through matched benchmark families. Harnesses and reported settings remain attached to the evidence. These results describe the evaluated configurations; they do not establish how a model will perform in every workflow.

View the Capability ranking →

Behind the ranking

273 entries with published results · 127 ranked · 0 provisional estimates.

Explore seven capabilities, the original results, and gaps in the evidence.

Coding ranking Reasoning ranking Agentic ranking Compare capabilities ↗Published benchmarks ↗Evidence coverage ↗

Cite this ranking

Include the model and effort setting, the selected profile, and the dated data downloads. A shared link uses the latest data; it does not freeze a ranking in time.

UnifyBench ranking
Data updated: 12 Sept 2026
Mode: custom; evidence: documented; configurations: best
Weights: v3;agentic:100,hard_reasoning:0,coding:0,human_pref:0,knowledge:0,multimodal:0,long_context:0
https://unifybench.ai/rankings/agentic

Catalog data · Effort measurements · Configuration scoring rules · Methodology

Catalog dated 12 Sept 2026. Effort evidence collected 12 Sept 2026. Collection dates are not evaluation dates.

What does the score mean?

The adjusted comparison score estimates performance against the same reference panel from matched published comparisons. Capability pools matched comparisons across families and checks aggregate breadth and independent corroboration. Custom profiles require support in every selected capability. Missing results never count as losses.

Family details show observed win shares against available reference peers. Those descriptive values are supporting evidence, not adjusted capability scores. Provider reports and independent results remain labeled. Matching reported settings do not establish equal inference effort, and score differences do not establish statistical significance.

Read the full methodology →