Frontier LLMs: $0.15 or more a review call

Models: 12; 2 scored
Cheapest call: $0.1532; Claude Opus 5.5
Dearest call: $1.2213; GPT-5.5 Pro
Best bugs caught: 90%; Claude Opus 5.5
Newest: 2026-09-22; Claude Opus 5.5

Pricing

Frontier-tier models: list prices per 1M tokens, one review call and reading a pull request.
				Cache write / 1M
Claude Opus 5.5Anthropic	$4.00	$20.00	$0.20	$5.00	$0.1532	$0.1048
GPT-5.6 SolOpenAIAbove 272K in: $8.00 / $30.00	$4.00	$20.00	$0.40	–	$0.1532	$0.1048
Claude Opus 5Anthropic	$5.00	$25.00	$0.50	$6.25	$0.1915	$0.131
Claude Opus 4.8Anthropic	$5.00	$25.00	$0.50	$6.25	$0.1915	$0.131
Claude Opus 4.7Anthropic	$5.00	$25.00	$0.50	$6.25	$0.1915	$0.131
Claude Opus 4.6Anthropic	$5.00	$25.00	$0.50	$6.25	$0.1915	$0.131
GPT-5.5OpenAIAbove 272K in: $10.00 / $45.00	$5.00	$30.00	$0.50	–	$0.2035	$0.131
Claude Fable 5.1Anthropic	$10.00	$50.00	$0.25	$12.50	$0.3829	$0.2621
Claude Fable 5Anthropic	$10.00	$50.00	$1.00	$12.50	$0.3829	$0.2621
GPT-6 AstraOpenAIAbove 272K in: $20.00 / $75.00	$10.00	$50.00	$1.00	–	$0.3829	$0.2621
GPT-5.5 ProOpenAI · Responses API onlyAbove 272K in: $60.00 / $270.00	$30.00	$180.00	–	–	$1.2213	$0.7862
GPT-5.4 ProOpenAI · Responses API onlyAbove 272K in: $60.00 / $270.00	$30.00	$180.00	–	–	$1.2213	$0.7862

Benchmark

Benchmark results
Compare
	1=	Claude Opus 5.5Anthropic · Frontier	90%	93%	98%	1.7	$0.1525	$0.1921	33s
	1=	GPT-5.6 SolOpenAI · Frontier	72%	72%	98%	1.2	$0.1127	$0.1732	50s

By language and bug kind

Share of known bugs each model caught, by language
Model	All	TypeScript6 bugs	Go4 bugs	Java4 bugs	Rust4 bugs	C#3 bugs	JavaScript3 bugs	Kotlin3 bugs	PHP3 bugs	Python3 bugs	Ruby3 bugs	Swift3 bugs	C++2 bugs	C1 bug
Claude Opus 5.5	90%	100%	50%	78%	100%	100%	100%	67%	100%	100%	89%	100%	100%	100%
GPT-5.6 Sol	72%	67%	50%	69%	75%	56%	100%	67%	100%	100%	67%	100%	50%	0%

Quality against cost

Higher catches more · further left is cheaper

AnthropicOpenAIFaded: models outside this list95% range: the line through each point

Other benchmarks

Outside benchmarks by model
Model	Arena WebDev	Arena Agent	Epoch Capabilities Index
Claude Fable 5.1	1741Max · range 1732 to 1751	0.140Max · range 0.120 to 0.160	165range 162 to 169
Claude Fable 5	1625High · range 1619 to 1631	0.082High · range 0.068 to 0.096	162range 160 to 166
Claude Opus 5.5	1807Max · range 1793 to 1822	0.189Max · range 0.161 to 0.216	167range 164 to 171
Claude Opus 5	1689Max · range 1683 to 1696	0.087Max · range 0.070 to 0.104	163range 160 to 166
Claude Opus 4.8	1556High · range 1550 to 1561	–	158range 156 to 161
Claude Opus 4.7	1558High · range 1552 to 1563	–	156range 155 to 159
Claude Opus 4.6	1547High · range 1541 to 1552	–	155range 154 to 157
GPT-6 Astra	1777Max · range 1767 to 1788	0.141Max · range 0.118 to 0.164	166range 163 to 171
GPT-5.6 Sol	–	0.071xHigh · range 0.057 to 0.085	162range 160 to 165
GPT-5.5 Pro	–	–	162range 159 to 165
GPT-5.4 Pro	–	–	159range 157 to 162
GPT-5.5	–	0.031xHigh · range 0.019 to 0.044	159range 156 to 162

Source: LMArena CC BY 4.0, Oct 16, 2026 · Epoch AI CC BY 4.0, Oct 7, 2026

Updated Oct 16, 2026All prices pricing.json