What is the best AI model in this snapshot?

The first row and Best overall card show the current intelligence-first leader. The best choice can change once price, latency, hosting and task-specific evaluation enter the decision.

What is the strongest open model?

The Best open model card is recalculated from models with verified downloadable weights, using the same intelligence-first ordering as the main table.

Which model has the best price-to-performance?

The Best value card is recalculated from the published value formula whenever verified intelligence, speed or base token pricing changes.

Does the largest context window mean the best result?

No. Context capacity only describes how much can be submitted. Retrieval design, instruction quality, attention behaviour and evaluation still determine whether that context produces a useful answer.

How often is the leaderboard updated?

A scheduled pipeline revalidates the dataset weekly and after major releases. Failed or conflicting source checks keep the last verified snapshot online instead of publishing partial data.

How are models ranked?

The default order uses the public intelligence signal. Speed, price, context, openness and a separate value formula remain visible so teams can challenge the default order.

Is the leaderboard vendor-independent?

Tenten does not sell model tokens through this page. Public sources are linked, assumptions are shown, and we recommend a task-specific evaluation before any production decision.

AI · LEADERBOARD

最強ではなく、仕事に合うモデルを選ぶ。

今導入できるAIモデルを、知能、速度、コスト、コンテキスト、運用条件で比較します。

Discuss your model choice View use cases

Snapshot: Aug 25, 2026, 10:41 AM UTC+816 models compared

Independent editorial snapshot · not live platform telemetry

Dated editorial snapshot

Review target: weekly and after major releases. Rankings reflect this dated snapshot. OpenRouter usage below has its own source timestamp.

AI · LEADERBOARDSnapshot

01Claude Opus 5Anthropic63.1
02Claude Fable 5Anthropic62.1
03GPT-5.6 SolOpenAI60.9
04Grok 4.6SpaceXAI60.9
05Kimi K3Kimi59.7

Intelligence16 models compared

The short answer

Three leaders, three different decisions.

Best overall

Closed model

Anthropic

Intelligence: 63.1
Speed: 58.8tok/s
Input / output: $5 / $25
Context: 1M

Best fit

State-of-the-art reasoning and complex agentic workflows

Best open model

Open weights

Kimi

Intelligence: 59.7
Speed: 35tok/s
Input / output: $3 / $15
Context: 1.05M

Best fit

Sovereign, open-weight deployments needing maximum intelligence

Best value

Open weights

DeepSeek

Intelligence: 51.8
Speed: 115tok/s
Input / output: $0.14 / $0.28
Context: 1.31M

Best fit

Cost-sensitive open inference and coding workloads

Model explorer

Rank, filter and compare.

No single score tells the whole story. Narrow the list, inspect the trade-offs, then compare the finalists side by side.

16models

						Best for
#1	63.1	58.8 tok/s	18.1	$5 / $25input / output	1M	State-of-the-art reasoning and complex agentic workflows
#2	62.1	71 tok/s	8.6	$10 / $50input / output	1M	High-reliability enterprise work with fallback safety
#3	60.9	73.6 tok/s	40.3	$2 / $10input / output	1.1M	Enterprise reasoning and large-context problem solving
#4	60.9	61.9 tok/s	64.5	$2 / $6input / output	500K	High-intelligence proprietary reasoning at competitive token pricing
#5	59.7	35 tok/s	25.1	$3 / $15input / output	1.0M	Sovereign, open-weight deployments needing maximum intelligence
#6	58.1	23 tok/s	42.0	$2 / $6input / output	1M	High-intelligence multilingual work and long-form translation
#7	56.8	N/A	N/A	$1.25 / $4.25input / output	1.0M	Massive-context analysis and structured data extraction
#8	56.6	121.5 tok/s	26.2	$2 / $12input / output	1.1M	Fast enterprise reasoning with balanced economics
#9	55.8	58 tok/s	47.4	$2 / $6input / output	500K	Fast reasoning where live information matters
#10	55.3	69 tok/s	28.7	$2 / $10input / output	1M	Balanced agents, coding and long-context workflows
#11	53.2	71.7 tok/s	70.9	$1.12 / $3.37input / output	1.0M	Low-cost open-weight reasoning and coding workloads
#12	52.6	152.8 tok/s	61.4	$1.19 / $3.74input / output	1.0M	Sovereign deployments that still need strong reasoning
#13	52.3	140.7 tok/s	196.4	$0.20 / $1.2input / output	1.1M	High-volume customer support with low latency and controlled cost
#14	51.8	115 tok/s	732.9	$0.14 / $0.28input / output	1.3M	Cost-sensitive open inference and coding workloads
#15	51.6	221 tok/s	59.0	$0.75 / $3.75input / output	1.0M	Real-time support, translation and multimodal intake
#16	38.0	82 tok/s	172.4	$0.14 / $0.28input / output	1.1M	Cost-sensitive extraction and high-volume structured work

Intelligence: 63.1
Speed: 58.8 tok/s
Price / 1M: $5 / $25
Context: 1M

Best fit

State-of-the-art reasoning and complex agentic workflows

Intelligence: 62.1
Speed: 71 tok/s
Price / 1M: $10 / $50
Context: 1M

Best fit

High-reliability enterprise work with fallback safety

Intelligence: 60.9
Speed: 73.6 tok/s
Price / 1M: $2 / $10
Context: 1.1M

Best fit

Enterprise reasoning and large-context problem solving

Intelligence: 60.9
Speed: 61.9 tok/s
Price / 1M: $2 / $6
Context: 500K

Best fit

High-intelligence proprietary reasoning at competitive token pricing

Intelligence: 59.7
Speed: 35 tok/s
Price / 1M: $3 / $15
Context: 1.0M

Best fit

Sovereign, open-weight deployments needing maximum intelligence

Intelligence: 58.1
Speed: 23 tok/s
Price / 1M: $2 / $6
Context: 1M

Best fit

High-intelligence multilingual work and long-form translation

Intelligence: 56.8
Speed: N/A
Price / 1M: $1.25 / $4.25
Context: 1.0M

Best fit

Massive-context analysis and structured data extraction

Intelligence: 56.6
Speed: 121.5 tok/s
Price / 1M: $2 / $12
Context: 1.1M

Best fit

Fast enterprise reasoning with balanced economics

Intelligence: 55.8
Speed: 58 tok/s
Price / 1M: $2 / $6
Context: 500K

Best fit

Fast reasoning where live information matters

Intelligence: 55.3
Speed: 69 tok/s
Price / 1M: $2 / $10
Context: 1M

Best fit

Balanced agents, coding and long-context workflows

Intelligence: 53.2
Speed: 71.7 tok/s
Price / 1M: $1.12 / $3.37
Context: 1.0M

Best fit

Low-cost open-weight reasoning and coding workloads

Intelligence: 52.6
Speed: 152.8 tok/s
Price / 1M: $1.19 / $3.74
Context: 1.0M

Best fit

Sovereign deployments that still need strong reasoning

Intelligence: 52.3
Speed: 140.7 tok/s
Price / 1M: $0.20 / $1.2
Context: 1.1M

Best fit

High-volume customer support with low latency and controlled cost

Intelligence: 51.8
Speed: 115 tok/s
Price / 1M: $0.14 / $0.28
Context: 1.3M

Best fit

Cost-sensitive open inference and coding workloads

Intelligence: 51.6
Speed: 221 tok/s
Price / 1M: $0.75 / $3.75
Context: 1.0M

Best fit

Real-time support, translation and multimodal intake

Intelligence: 38.0
Speed: 82 tok/s
Price / 1M: $0.14 / $0.28
Context: 1.1M

Best fit

Cost-sensitive extraction and high-volume structured work

OpenRouter usage

OpenRouter LLM usage ranking

Top models by prompt, completion, and reasoning tokens processed on OpenRouter. Showing the first 20 only.

Top 20OpenRouter data · cached hourlyUpstream · Oct 6, 2026

1.DeepSeek V4.1 Flash (batch)by DeepSeek29.6T tokens34%
2.GLM 5.3 Flash (batch)by Z.ai10.2T tokens12%
3.MiMo-V2.6-Flashby Xiaomi9.87T tokens20%
4.GPT-6 Luna (batch)by OpenAI6.44T tokens47%
5.Hy4 previewby Tencent6.42T tokens20%
6.DeepSeek V4 Flash 0731by DeepSeek5.79T tokens20%
7.Nemotron 3 Ultra (free)by Nvidia5.68T tokens3%
8.DeepSeek V4 Flash 0423by DeepSeek3.23T tokens1%
9.GLM 5.3 (batch)by Z.ai3.06T tokens9%
10.Claude Opus 5.5 (batch)by Anthropic2.93T tokens73%
11.GPT-5.6 Luna (batch)by OpenAI2.36T tokens72%
12.Gemini 3.8 Flash (batch)by Google2.17T tokens6%
13.GPT-5.6 Sol (batch)by OpenAI1.77T tokens8%
14.Kimi K3 (batch)by Moonshot AI1.7T tokens14%
15.Hy3by Tencent1.68T tokens31%
16.Muse Spark 1.3 Contributorby Meta1.43T tokens7%
17.GLM 5.2by Z.ai1.42T tokens3%
18.GPT-6 Astra (batch)by OpenAI1.38T tokens50%
19.MiMo-V2.6-Proby Xiaomi1.22T tokens1%
20.Claude Sonnet 5.5 (batch)by Anthropic1.19T tokens955%

From benchmark to decision

Which model for which job?

Nine recurring production scenarios, with the trade-offs that matter before a team commits to a model.

Selected use case

01 / 09

Decision view

Customer support

Why it fits

01High measured throughput
02Low per-token cost
03Strong intelligence for support

How the ranking works

Transparent enough to disagree with.

The ranking combines public benchmark signals with price, measured throughput, context and deployment constraints. It is a decision aid, not a universal definition of model quality.

Dataset snapshot: 25 Aug 2026 · 10:41 TST
Review cadence: Revalidated weekly and after major releases

Value score

(intelligence − 25)² × min(1, speed ÷ 30) ÷ (input cost + output cost × 3)

Output is weighted ×3 to reflect output-heavy production use; intelligence is squared so cheap but weak models do not dominate.

Intelligence: Public benchmark and reasoning signals form the default order.
Speed: Output tokens per second; provider and region still affect real latency.
Cost: Public input and output price per million tokens, before enterprise discounts.
Deployment: Context, weight access and hosting constraints determine production fit.

Public references

Prices and performance can change without notice. Validate the finalists against your own prompts, infrastructure and risk requirements before procurement.

FAQ

Questions worth asking before the benchmark wins the argument.

TENTEN AI · MODEL EVALUATION

The leaderboard is the start

Bring the workflow. We’ll test the shortlist.

A model only becomes a good decision after it survives your documents, permissions, latency budget and failure cases.

Discuss a workflow

最強ではなく、仕事に合うモデルを選ぶ。

Three leaders, three different decisions.

Rank, filter and compare.

Claude Opus 5

Claude Fable 5

GPT-5.6 Sol

Grok 4.6

Kimi K3

Qwen3.8 Max

Muse Spark 1.2

GPT-5.6 Terra

Grok 4.5

Claude Sonnet 5

DeepSeek V4 Pro 0813

GLM-5.2

GPT-5.6 Luna

DeepSeek V4 Flash 0731

Gemini 3.6 Flash

MiMo-V2.5

OpenRouter LLM usage ranking

Which model for which job?

Customer support

Transparent enough to disagree with.

Questions worth asking before the benchmark wins the argument.

01What is the best AI model in this snapshot?

02What is the strongest open model?

03Which model has the best price-to-performance?

04Does the largest context window mean the best result?

05How often is the leaderboard updated?

06How are models ranked?

07Is the leaderboard vendor-independent?

Bring the workflow. We’ll test the shortlist.

Rank, filter and compare.

Claude Opus 5

Claude Fable 5

GPT-5.6 Sol

Grok 4.6

Kimi K3

Qwen3.8 Max

Muse Spark 1.2

GPT-5.6 Terra

Grok 4.5

Claude Sonnet 5

DeepSeek V4 Pro 0813

GLM-5.2

GPT-5.6 Luna

DeepSeek V4 Flash 0731

Gemini 3.6 Flash

MiMo-V2.5