Skip to content
[ Frontier / cloud models ]

The big models, honestly priced

Every frontier API model on one board: live per-token pricing, verified benchmark scores with sources, context windows — and for each one, what it would actually take to run something comparable on your own hardware.

Compare two models →

Current leaders

Overall best
65

Intelligence index

Claude Fable 5

Best coding
95%

SWE-bench Verified

Claude Fable 5

Cheapest capable
$0.13

Blended / 1M tokens

DeepSeek V4 Flash

Largest context
10M

Token window

Llama 4 Scout

Model catalogue

25 models tracked
Rank by
Price / 1M
25 / 25 models
AnthropicAPI only

Claude Fable 5

Anthropic's Mythos-class flagship — state-of-the-art on nearly every tested benchmark.

Intelligence65
Coding · SWE-bench95
Context
1M
In / 1M
$10
Out / 1M
$50
AnthropicAPI only

Claude Opus 4.8

Anthropic's flagship — top of the SWE-bench Verified leaderboard.

Intelligence61.4
Coding · SWE-bench88.6
Context
1M
In / 1M
$5
Out / 1M
$25
OpenAIAPI only

GPT-5.5

OpenAI's flagship — state of the art on terminal/agentic benchmarks.

Intelligence60
Coding · SWE-bench82.6
Context
1.05M
In / 1M
$5
Out / 1M
$30
AnthropicAPI only

Claude Opus 4.7

Previous Anthropic flagship — still elite, especially on science reasoning.

Intelligence57
Coding · SWE-bench82
Context
1M
In / 1M
$5
Out / 1M
$25
GoogleAPI only

Gemini 3.1 Pro (Preview)

Google's reasoning flagship — top of the GPQA Diamond leaderboard.

Intelligence57
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$2
Out / 1M
$12
OpenAIAPI only

GPT-5.4

Previous OpenAI flagship — now the value pick of the GPT-5 line at half price.

Intelligence57
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$2.50
Out / 1M
$15
AlibabaAPI only

Qwen3.7 Max

Alibaba's top tier — 1M context, strong multilingual.

Intelligence57
Coding · SWE-benchn/a
Context
1M
In / 1M
$1.25
Out / 1M
$3.75
GoogleAPI only

Gemini 3.5 Flash

Google's speed/quality sweet spot — near-flagship scores at Flash pricing.

Intelligence55
Coding · SWE-bench78.8
Context
1.05M
In / 1M
$1.50
Out / 1M
$9
MiniMaxOpen weights

MiniMax M3

Open-weights with video input and a 512K output window.

Intelligence55
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$0.30
Out / 1M
$1.20
Moonshot AIOpen weights

Kimi K2.6

The open-source favourite for agentic coding.

Intelligence54
Coding · SWE-bench80.2
Context
262K
In / 1M
$0.67
Out / 1M
$3.39
xAIAPI only

Grok 4.3

xAI's newer mainline — 1M context, same low price.

Intelligence53
Coding · SWE-benchn/a
Context
1M
In / 1M
$1.25
Out / 1M
$2.50
AlibabaAPI only

Qwen3.7 Plus

The value Qwen — 1M context with vision at $0.32/$1.28.

Intelligence53
Coding · SWE-benchn/a
Context
1M
In / 1M
$0.32
Out / 1M
$1.28
AnthropicAPI only

Claude Sonnet 4.6

The default workhorse Claude — 60% of Opus quality questions at 60% of the price.

Intelligence52
Coding · SWE-benchn/a
Context
1M
In / 1M
$3
Out / 1M
$15
DeepSeekOpen weights

DeepSeek V4 Pro

The open-weights flagship — AA index 52 at one-tenth of flagship prices.

Intelligence52
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$0.43
Out / 1M
$0.87
Z.aiOpen weights

GLM-5.1

Zhipu's open flagship — a SWE-bench official-leaderboard regular.

Intelligence51
Coding · SWE-bench72.8
Context
203K
In / 1M
$0.98
Out / 1M
$3.08
xAIAPI only

Grok 4.20

2M-token context and aggressive pricing — xAI's speed-focused reasoner.

Intelligence49.3
Coding · SWE-benchn/a
Context
2M
In / 1M
$1.25
Out / 1M
$2.50
OpenAIAPI only

GPT-5.4 Mini

Mid-size OpenAI tier — solid quality at $0.75/$4.50.

Intelligence49
Coding · SWE-benchn/a
Context
400K
In / 1M
$0.75
Out / 1M
$4.50
DeepSeekOpen weights

DeepSeek V4 Flash

Absurdly cheap open-weights workhorse — $0.10/$0.20 per 1M.

Intelligence47
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$0.10
Out / 1M
$0.20
OpenAIAPI only

GPT-5.4 Nano

OpenAI's cheapest tier — $0.20/$1.25 for bulk workloads.

Intelligence44
Coding · SWE-benchn/a
Context
400K
In / 1M
$0.20
Out / 1M
$1.25
AnthropicAPI only

Claude Haiku 4.5

Anthropic's fast/cheap tier for high-volume calls.

Intelligence37
Coding · SWE-benchn/a
Context
200K
In / 1M
$1
Out / 1M
$5
GoogleAPI only

Gemini 3.1 Flash Lite

Google's bulk tier — 1M multimodal context at $0.25/$1.50.

Intelligence34
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$0.25
Out / 1M
$1.50
Mistral AIOpen weights

Mistral Large 2512

The European flagship — open-weights, EU-hosted options.

Intelligence23
Coding · SWE-benchn/a
Context
262K
In / 1M
$0.50
Out / 1M
$1.50
MetaOpen weights

Llama 4 Maverick

Meta's open workhorse — huge ecosystem, runs everywhere.

Intelligence18
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$0.20
Out / 1M
$0.80
MetaOpen weights

Llama 4 Scout

10M-token context — the longest window of any model, open weights.

Intelligence14
Coding · SWE-benchn/a
Context
10M
In / 1M
$0.10
Out / 1M
$0.30
OpenAIAPI only

GPT-5.5 Pro

GPT-5.5 with extended parallel compute — 6× the price for the hardest problems.

Intelligencen/a
Coding · SWE-benchn/a
Context
1.05M
In / 1M
$30
Out / 1M
$180

How to read these numbers. Pricing and context windows come from a live OpenRouter snapshot and are re-verified weekly. Benchmark conventions: Intelligence is a composite index (Artificial-Analysis style), Coding is SWE-bench Verified %, Reasoning is GPQA Diamond %. Every score cites its source on the model's datasheet — and where a number isn't independently verified, we show n/a rather than guessing. Blended price assumes a 3:1 input:output token mix.