LLM leaderboard, July 2026

LLM Leaderboard & AI Model Benchmarks — July 2026

Compare frontier AI models by quality, cost, and context. 79 ranked and 35 verified models among 281 tracked LLMs — 296 benchmarks, real pricing, and runtime data in one place.

Data refreshed: July 11, 2026
Verified coverage: 35 ranked models
Ranking method: No interpolation

The LLM leaderboard — July 2026 rankings

Benchmarks, pricing, runtime signals, and context window in one table. Filter state syncs to the URL so every view is shareable. Provisional-ranked mode includes source-unverified non-generated benchmark evidence.

Methodology Providers AI Race

CSV JSON


1 Claude Fable 5 Anthropic	Anthropic	Closed	Current	Reasoning	1M+	$10.00 / $50.00	Not measured	Not measured	91	99	94	—	80	89	100	88	85	1508.51
2 Claude Mythos 5 Anthropic	Anthropic	Closed	Current	Reasoning	1M+	$10.00 / $50.00	Not measured	Not measured	89	100	91	—	93	88	100	86	82	—
3 Gemini 3.1 Pro Google	Google	Closed	Current	Standard	1M	$2.00 / $12.00	109	29.71s	88	77	89	96	83	91	100	88	62	1485.8
4 Gemini 3 Pro Deep Think Google	Google	Closed	Current	Reasoning	2M	Not listed	Not measured	Not measured	~87	92	73	89	95	84	78	79	—	1486.39
5 GPT-5.4 OpenAI	OpenAI	Closed	Superseded	Reasoning	1.05M	$2.50 / $15.00	74	151.79s	85	88	86	88	60	97	100	93	71	1467.22
6 Claude Opus 4.8 Anthropic	Anthropic	Closed	Current	Reasoning	1M	$5.00 / $25.00	Not measured	Not measured	85	94	90	—	66	88	—	—	74	1476.51
7 Claude Opus 4.6 Anthropic	Anthropic	Closed	Superseded	Standard	1M	$5.00 / $25.00	40	1.78s	83	79	82	88	75	92	100	91	54	1498.93
8 Qwen3.7 Max Alibaba	Alibaba	Closed	Current	Reasoning	1M	Not listed	Not measured	Not measured	83	76	88	79	—	79	84	90	83	1474.85
9 GPT-5.3 Codex OpenAI	OpenAI	Closed	Current	Reasoning	400K	$1.75 / $14.00	79	88.26s	~82	78	84	88	91	89	95	88	—	1416
10 GLM-5.2 Z.AI Self-host	Z.AI	Open	Current	Reasoning	1M	$1.40 / $4.40	Not measured	Not measured	81	85	79	—	—	85	—	—	81	—
11 Gemini 3.5 Flash Google	Google	Closed	Current	Reasoning	1M	$1.50 / $9.00	284.2	18.55s	80	87	75	73	76	72	—	76	62	1479.46
12 Gemini 3 Pro Google	Google	Closed	Established	Standard	2M	$2.00 / $12.00	109	32.65s	80	82	68	82	80	77	78	80	60	1485.77
13 GPT-5.5 OpenAI	OpenAI	Closed	Superseded	Reasoning	1M	$5.00 / $30.00	Not measured	Not measured	78	96	73	82	58	82	—	—	70	1474.62
14 DeepSeek V4 Pro (Max) DeepSeek Self-host	DeepSeek	Open	Current	Reasoning	1M	$1.74 / $3.48	Not measured	Not measured	78	78	83	—	—	72	—	—	82	—
15 Qwen3.7 Plus Alibaba	Alibaba	Closed	Current	Reasoning	1M	Not listed	Not measured	Not measured	78	69	81	78	73	73	80	89	79	1462.53
16 GLM-5 (Reasoning) Z.AI Self-host	Z.AI	Open	Current	Reasoning	200K	$1.00 / $3.20	Not measured	Not measured	~77	88	69	82	71	78	78	76	33	1455.62
17 Claude Sonnet 4.6 Anthropic	Anthropic	Closed	Superseded	Standard	200K	$3.00 / $15.00	44	1.48s	77	70	76	81	84	84	90	80	49	1472.16
18 Claude Opus 4.7 (Adaptive) Anthropic	Anthropic	Closed	Superseded	Reasoning	1M	$5.00 / $25.00	Not measured	Not measured	76	83	83	76	50	85	—	—	—	1501.97
19 GPT-5.2 OpenAI	OpenAI	Closed	Established	Reasoning	400K	$1.75 / $14.00	73	130.34s	75	53	74	83	81	88	95	79	57	1434.98
20 DeepSeek V4 Pro (High) DeepSeek Self-host	DeepSeek	Open	Current	Reasoning	1M	$1.74 / $3.48	Not measured	Not measured	74	69	85	—	—	68	—	—	81	—
21 Kimi K2.6 Moonshot AI Self-host	Moonshot AI	Open	Superseded	Reasoning	256K	$0.95 / $4.00	Not measured	Not measured	74	80	88	—	68	63	—	—	72	1461.37
22 Kimi K2.5 (Reasoning) Moonshot AI Self-host	Moonshot AI	Closed	Current	Reasoning	128K	$0.60 / $3.00	Not measured	Not measured	73	58	89	65	67	72	81	89	—	1447
23 GPT-5.1 OpenAI	OpenAI	Closed	Established	Reasoning	200K	$1.25 / $10.00	111	57.47s	~73	72	68	66	92	77	84	73	48	1438.78
24 DeepSeek V4 Flash (Max) DeepSeek Self-host	DeepSeek	Open	Current	Reasoning	1M	$0.14 / $0.28	Not measured	Not measured	72	66	80	—	—	65	—	—	81	—
25 GPT-5 (high) OpenAI	OpenAI	Closed	Established	Reasoning	128K	$1.25 / $10.00	83	36.28s	~72	73	64	72	89	75	78	76	—	1433.63

Showing 25 of 281

Reading the table

Use rank to shortlist, then inspect the evidence

Embed the live leaderboard

The BenchLM LLM leaderboard 2026 provisionally ranks 79+ models and tracks 281+ large language models side by side across 296 benchmarks — from SWE-bench and LiveCodeBench for coding to GPQA Diamond and MMLU-Pro for knowledge and reasoning. Whether you need the best AI models 2026 has to offer for agentic workflows, math, multilingual tasks, or instruction following, our AI benchmark comparison tables make it easy to see how GPT-5, Claude, Gemini, DeepSeek, Llama, and dozens of other frontier and open-source models stack up on both benchmarks and operator tradeoffs like price and context. The main leaderboard distinguishes provisional ranking from verified ranking so you can see which scores rest on exact-source coverage and which still rely on source-unverified public rows.

Decision-ready picks

Current models that clear the ranking, freshness, and evidence thresholds for each decision.

Qualification method

DecisionModelCurrent measureProvenance

Best open weightHighest-scoring evidence-qualified current open-weight model.GLM-5.2Z.AI81overall scoreVerified-ranked · 15 exact-source benchmarks
Best near-frontier value97% of the leading score at a lower output price.Gemini 3.1 ProGoogle$12output / 1M tokensProvisional-ranked · 19 public benchmark rows
Fastest measuredFastest externally measured model that also clears the ranking and evidence thresholds.Gemini 3.5 FlashGoogle284.2tokens / secArtificial Analysis · updated 2026-07-06
Largest useful contextLargest context among evidence-qualified current models retaining at least half of the leading score.Grok 4.20xAI2Mcontext windowProvisional-ranked · 16 public benchmark rows

Qualification method

Dataset journal

What changed in July 2026

Claude Fable 5 (Anthropic) holds the #1 spot on the leaderboard at 91.
11 model releases tracked in July 2026.
35 models now carry fully verified rankings.

Data verified July 11, 2026

Build a head-to-head comparison

Choose any two ranked models. The comparison opens with the decision, exact units, and evidence coverage.

versus

Compare these models

The evidence map

Explore benchmarks by category

Move from the overall ranking into the capability or workload that actually decides your model choice.

Agentic56Tools, browsers, and long-running work Coding39Repository tasks and code generation Multimodal51Documents, images, and grounded work Reasoning25Long-context and abstract reasoning Knowledge33Professional and scientific knowledge Instruction following4Constraint and format reliability Multilingual11Cross-language accuracy and coverage Math27Contest and frontier mathematics

Popular comparisons

GPT-5.6 Sol vs Claude Sonnet 5 GPT-5.6 Sol vs Claude Opus 4.8 GPT-5.6 Terra vs Gemini 3.5 Flash GPT-5.6 Luna vs Gemini 3.5 Flash Claude Sonnet 5 vs Claude Opus 4.8 Claude Opus 4.8 vs GPT-5.5 Claude Opus 4.8 vs Claude Opus 4.7 (Adaptive)Gemini 3.5 Flash vs GPT-5.5 Gemini 3.5 Flash vs Claude Opus 4.7

Tools and resources

LLM Pricing ComparisonCompare API costs across models Alternative FinderFind the best ChatGPT or Claude replacement Which LLM Should I Use?Get a personalized recommendation Cost CalculatorEstimate your monthly API spend Methodology & ExplainersHow BenchLM scores and curates benchmarks

Scoring methodology8 weighted categories, 29 ranking benchmarks, and 217 display-only records.Inspect method

Each model's overall score is a normalized weighted average of category averages. Within each category, benchmark results are normalized to a common scale and combined using weights that favor harder, less-saturated evaluations.

Confidence remains separate from score. It shows how much public benchmark evidence supports a result, while display-only benchmarks stay visible for context without receiving ranking weight.

Data comes from OpenBench, official model papers, and public leaderboards. External consensus signals are bounded calibration inputs and are not exposed in exported data.

Read the complete methodology

Category weights: 100%
Scored categories: 8
Ranking benchmarks: 29
Display-only: 217

Category and weightBenchmarks that affect the scoreCoverage

Agentic22%
Terminal-Bench 2.0 · BrowseComp · OSWorld-Verified
3 weighted · 53 context-only
Coding20%
SWE-bench Verified · SWE-Rebench · LiveCodeBench · SWE-bench Pro · SciCode
5 weighted · 34 context-only
Reasoning17%
MuSR · LongBench v2 · MRCRv2 · ARC-AGI-2
4 weighted · 21 context-only
Multimodal12%
MMMU-Pro · OfficeQA Pro · CharXiv
3 weighted · 48 context-only
Knowledge12%
GPQA · SuperGPQA · MMLU-Pro · HLE · FrontierScience · SimpleQA
6 weighted · 27 context-only
Multilingual7%
MMLU-ProX
1 weighted · 10 context-only
Instruction Following5%
IFEval · IFBench
2 weighted · 2 context-only
Math5%
AIME26 · HMMT Feb 2026 · FrontierMath v2 (Tiers 1-3) · FrontierMath v2 (Tier 4) · USAMO 2026
5 weighted · 22 context-only

LLM Leaderboard & AI Model Benchmarks — July 2026

The LLM leaderboard — July 2026 rankings

Use rank to shortlist, then inspect the evidence

Decision-ready picks

What changed in July 2026

Build a head-to-head comparison

Explore benchmarks by category

More ranking paths

Choose by workload

The AI models change fast. We track them for you.

Popular comparisons

Tools and resources