LLM Leaderboard & AI Model Benchmarks — October 2026

As of October 6, 2026, the BenchLM AI leaderboard ranks 214 models. Claude Opus 5.5 and GPT-6 Astra tie under the 90% interval rule. Open the evidence before choosing.

Search models and benchmarks

Current BenchAlign leaders

Snapshot October 6, 2026

Method

Claude Opus 5.5Supported
Anthropic01
Score86.3990% interval 80.5–92.3
Agentic
88
Coding
84
Reasoning
82
Multimodal
89
Knowledge
88
Multilingual
Not measured
Instruction following
Not measured
Math
Not measured
Supportedenough independent evidence
Details for Claude Opus 5.5
GPT-6 AstraSupported
OpenAI02
Score84.990% interval 79.9–90.0
Agentic
71
Coding
77
Reasoning
90
Multimodal
Not measured
Knowledge
85
Multilingual
Not measured
Instruction following
Not measured
Math
Not measured
Supportedenough independent evidence
Details for GPT-6 Astra
Claude Sonnet 5.5Supported
Anthropic03
Score83.990% interval 77.1–90.7
Agentic
68
Coding
85
Reasoning
79
Multimodal
Not measured
Knowledge
81
Multilingual
Not measured
Instruction following
Not measured
Math
Not measured
Supportedenough independent evidence
Details for Claude Sonnet 5.5

Price context: Claude Sonnet 5.5 retains 97% of the top score with an output price 50% lower.

Data refreshed: October 6, 2026
Supported: 83 models
Ranking method: BenchAlign v5.8

The LLM leaderboard — October 2026 rankings

Score, model size, context, price, and benchmark detail in one table. Filter state syncs to the URL so every view is shareable. Supported and Estimated labels show how much independent evidence backs each position.

Methodology Providers AI Race

CSV API/MCP

Supported means we have enough independent evidence to stand behind the row. Estimated means the rank is visible and the uncertainty is wider.

Supportedenough independent evidenceEstimatedwider uncertainty

Parameters show total billions; active counts appear when a source reports them. Size filters use total parameters: Tiny ≤4B, Small >4–<40B, Medium 40–150B, and Large >150B.

AG Agentic CO Coding RE Reasoning MM Multimodal & Grounded KN Knowledge ML Multilingual IF Instruction Following MA Math

LLM leaderboard table

#1 Claude Opus 5.5 AnthropicSupported	86.39	Not disclosed	1M	$4.00 / $20.00	1503.7	Closed	88	84	82	89	88	—	—	—	97	744.81s	Proprietary	Current	Reasoning
#2 GPT-6 Astra OpenAISupported	84.9	Not disclosed	1.05M	$10.00 / $50.00	1477.11	Closed	71	77	90	—	85	—	—	—	63	309.32s	Proprietary	Current	Reasoning
#3 Claude Sonnet 5.5 AnthropicSupported	83.9	Not disclosed	1M	$2.00 / $10.00	1471.01	Closed	68	85	79	—	81	—	—	—	128	441.91s	Proprietary	Current	Reasoning
#4 Gemini 4 Argon GoogleEstimated	81.81	Not disclosed	—	$2.00 / $10.00	1525.22	Closed	75	73	—	—	78	—	—	—	Not measured	Not measured	Proprietary	Current	Reasoning
#5 Claude Fable 5.1 AnthropicSupported	81.76	Not disclosed	1M	$10.00 / $50.00	1501.14	Closed	79	78	81	—	85	—	—	—	62	259.51s	Proprietary	Current	Reasoning
#6 GPT-6.1 Sol OpenAIEstimated	81.46	Not disclosed	1.05M	$2.00 / $10.00	1483.21	Closed	72	71	87	—	83	—	—	—	Not measured	Not measured	Proprietary	Current	Reasoning
#7 Claude Opus 5 AnthropicSupported	79.33	Not disclosed	—	$5.00 / $25.00	1489.49	Closed	78	72	77	89	80	—	—	—	57	57.42s	Proprietary	Superseded	Reasoning
#8 Claude Fable 5 AnthropicSupported	78.82	Not disclosed	1M+	$10.00 / $50.00	1504.27	Closed	75	72	—	—	81	—	76	—	65	94.23s	Proprietary	Superseded	Reasoning
#9 GPT-5.6 Sol OpenAISupported	77.95	Not disclosed	1.05M	$4.00 / $20.00	1483.94	Closed	68	69	73	89	78	—	88	—	97	99.99s	Proprietary	Superseded	Reasoning
#10 GPT-6 Sol OpenAISupported	75.19	Not disclosed	1.05M	$2.00 / $10.00	1457.21	Closed	59	67	70	—	77	—	—	—	Not measured	Not measured	Proprietary	Superseded	Reasoning
#11 MiMo-V2.6-Pro XiaomiEstimated Self-host	74.17	Not disclosed	1M	$0.43 / $0.87	1480.08	Open	67	64	—	—	68	—	—	—	47	48.45s	Open Weight	Current	Reasoning
#12 Gemini 3.8 Flash GoogleSupported	73.06	Not disclosed	1M	$0.75 / $3.75	1494.77	Closed	65	63	71	84	72	—	—	—	242	16.76s	Proprietary	Current	Reasoning
#13 GPT-5.5 Pro OpenAIEstimated	72.9	Not disclosed	1M	$30.00 / $180.00	—	Closed	—	—	—	—	—	—	—	—	Not measured	Not measured	Proprietary	Current	Reasoning
#14 GPT-5.6 Terra OpenAISupported	72.33	Not disclosed	1.05M	$2.00 / $12.00	1465.48	Closed	59	64	66	78	70	—	86	—	Not measured	Not measured	Proprietary	Current	Reasoning
#15 Kimi K3 Moonshot AISupported	70.68	Not disclosed	1.05M	$3.00 / $15.00	1488.36	Pending	70	61	66	89	67	—	—	—	52	43.06s	Pending	Current	Reasoning
#16 Qwen3.8 Max AlibabaSupported Self-host	70.49	Not disclosed	1M	Not listed	1481.62	Open	65	54	—	88	66	—	90	—	37	56.22s	Open Weight	Current	Reasoning
#17 Step 5 Preview StepFunEstimated	70.26	Not disclosed	1M	$1.00 / $2.70	1453.68	Pending	63	61	—	61	67	—	—	—	87	25.73s	Pending	Current	Reasoning
#18 Claude Opus 4.8 AnthropicSupported	69.11	Not disclosed	1M	$5.00 / $25.00	1474.54	Closed	61	61	59	88	69	—	74	65	64	51.14s	Proprietary	Superseded	Reasoning
#19 GPT-5.4 Pro OpenAIEstimated	68.9	Not disclosed	1.05M	$30.00 / $180.00	1475.24	Closed	—	—	—	—	—	—	—	—	74	151.79s	Proprietary	Established	Reasoning
#20 GLM-5.3 Z.AISupported Self-host	68.68	753.33B	1M	$0.00 / $0.00	1478.5	Open	68	58	77	—	61	—	—	—	75	29.67s	Open Weight	Current	Reasoning
#21 Claude Opus 4.7 (Adaptive) AnthropicEstimated	67.96	Not disclosed	1M	$5.00 / $25.00	1502.21	Closed	59	56	—	50	64	—	—	—	Not measured	Not measured	Proprietary	Superseded	Reasoning
#22 Grok 4.6 xAISupported	67.95	Not disclosed	500K	$2.00 / $6.00	1453.83	Closed	68	60	58	—	68	—	—	—	73	28.15s	Proprietary	Superseded	Reasoning
#23 DeepSeek V4.1 Flash DeepSeekSupported Self-host	67.81	Not disclosed	1M	$0.30 / $1.20	1474.44	Open	61	62	—	—	64	—	—	—	222	10.09s	Open Weight	Current	Reasoning
#24 GPT-5.5 OpenAISupported	67.78	Not disclosed	1M	$5.00 / $30.00	1481.44	Closed	61	63	67	71	69	—	92	—	104	43.82s	Proprietary	Superseded	Reasoning
#25 Gemini 3.7 Flash GoogleSupported	67.6	Not disclosed	1M	$0.75 / $3.75	1487.68	Closed	58	59	—	84	70	—	—	—	284	10.62s	Proprietary	Superseded	Reasoning

Showing 25 of 882 models · 214 ranked

Source-linked · the table lists all 882 tracked models; 214 carry a BenchAlign rank

How the score is built

What the table covers

The BenchLM LLM leaderboard 2026 ranks 214+ models and tracks 882+ large language models side by side across 613 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. Supported and Estimated labels separate evidence strength from score, so incomplete reporting does not automatically remove a model from the comparison.

Reading the table

Embed the live leaderboard

What changed in October 2026

Claude Opus 5.5 (Anthropic) holds the #1 spot on the BenchAlign leaderboard at 86.39.
10 model releases tracked in October 2026.
83 Supported and 131 Estimated models appear in the public leaderboard.

Dataset journalData verified October 6, 2026

Explore benchmarks by category

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

The evidence mapBenchmark directory

Radar · latest confirmed releases

3 source-linked updates

Every model on this board will change. Radar keeps the record: releases, price changes, retirements, and API changes, each with its source and its date, matched to the models you declare.

EmbeddingGemma 2 releasedLatest
Google · Oct 6, 2026
Source confirmed
Mistral Large 4 released
Mistral · Oct 6, 2026
Source confirmed
Beam released
Reflection AI · Oct 5, 2026
Source confirmed

Follow model changes Browse public updates

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

Decision-ready picks

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

DecisionModelCurrent measureProvenance

Best open weightHighest-scoring evidence-qualified current open-weight model.MiMo-V2.6-ProXiaomi74.17overall scoreEstimated · 2 source families
Best near-frontier value97% of the leading score at a lower output price.Claude Sonnet 5.5Anthropic$10output / 1M tokensSupported · 3 source families
Fastest measuredFastest externally measured model that also clears the ranking and evidence thresholds.Mercury 2.5Inception691tokens / secArtificial Analysis · updated 2026-10-06
Largest useful contextLargest context among evidence-qualified current models retaining at least half of the leading score.GPT-6 AstraOpenAI1.05Mcontext windowSupported · 3 source families

Current models that clear rank, freshness and evidence thresholds

Qualification method Qualification method

Popular comparisons

Gemini 4 Argon vs GPT-6 Astra Gemini 4 Argon vs Claude Opus 5.5 Gemini 4 Argon vs Gemini 3.8 Flash Claude Opus 5.5 vs Claude Sonnet 5.5 Claude Opus 5.5 vs GPT-6 Astra Claude Opus 5.5 vs GPT-6.1 Sol Claude Opus 5.5 vs GPT-6 Sol Claude Sonnet 5.5 vs GPT-6.1 Sol Claude Sonnet 5.5 vs GPT-6 Sol

Tools and resources

AI Prompt WriterTurn the work you need into a testable prompt Prompt OptimizerGet closer to usable output with a clearer prompt 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 benchmarks are scored and curated

How we work

A ranking is a shortlist.

Rank orders the evidence; it does not make the decision. Every row here opens to its sources so you can disagree with us.

We do not invent scores to fill a table.

A model without a public, independent row stays unranked, and says so. Vendor-run numbers are shown, labeled, and kept out of the rank.

When we were wrong, the fix is on the page.

Scoring bugs, missing coverage, and re-weightings are dated and linked, because a ranking you cannot audit is a launch post with a table.

The desk

Editorial policy

The AI race, read as a record instead of a spectacle

Track which release led its month, how provider depth changed, and whether the crown moved.

Explore the release timeline

Months: 6
Releases: 219
Lead changes: 5

Best ranked model released in Oct 2026

Mistral Large 4Mistral

Score53.69

RankProviderTop-three avg.Models

01Anthropic8421

02OpenAI81.436

03Google74.219

Six-month release record

Explore the release timeline

Scoring methodology8 weighted categories, 61 ranking benchmarks, and 312 display-only records.Inspect method

Category weights: 100%
Scored categories: 8
Ranking benchmarks: 61
Display-only: 312

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.

Benchmark scores come from official provider tables, model cards, release announcements, and benchmark-native public leaderboards. External consensus signals are bounded calibration inputs and are not exposed in exported data.

Read the complete methodology

Category and weightBenchmarks that affect the scoreCoverage

Agentic22%
Terminal-Bench 3.0 · Terminal-Bench 4.0 · AA AutomationBench · AA EnterpriseOps-Gym · AA Harvey LAB · AA Tau3 Banking · Terminal-Bench 2.0 · Terminal-Bench 2.1 · BrowseComp · HLE w/ tools · GDPval-AA · OSWorld-Verified · OSWorld 2.0 · JobBench · MCP Atlas · Toolathlon · Toolathlon-Verified · AutomationBench · Agents' Last Exam · DeepSearchQA · BFCL v4 · τ³-bench results · Terminal-Bench 2.1 (Vals)
23 weighted · 76 context-only
Coding20%
SWE-Rebench · SWE-bench Pro · VulcanBench v3 · FrontierCode 1.1 Main · SWE Multilingual · FrontierSWE v2 · SciCode · AA-SciCode · LiveCodeBench (Vals) · DeepSWE
10 weighted · 53 context-only
Reasoning17%
LongBench v2 · MRCRv2 · ARC-AGI-2 · ARC-AGI-3 · AA-LCR · CritPt
6 weighted · 26 context-only
Multimodal12%
MMMU-Pro · AA-MMMU-Pro · OfficeQA Pro · CharXiv
4 weighted · 65 context-only
Knowledge12%
GPQA · SuperGPQA · MMLU-Pro · HLE · AA-GPQA Diamond · AA-HLE · AA-Omniscience Accuracy · HLE w/o tools · GPQA Diamond (Vals) · MMLU-Pro (Vals)
10 weighted · 47 context-only
Multilingual7%
MMLU-ProX
1 weighted · 13 context-only
Instruction Following5%
IFBench · AA-IFBench
2 weighted · 3 context-only
Math5%
AIME26 · HMMT Feb 2026 · FrontierMath v2 (Tiers 1-3) · FrontierMath v2 (Tier 4) · USAMO 2026
5 weighted · 29 context-only

Scoring methodology · BenchAlign v5.8

Inspect method