LLM API pricing

CodePanel prices 76 current models from 15 makers at their list rates: $0.0341 to $10.00 per 1M input tokens and $0.30 to $50.00 per 1M output tokens. One review call (26.2K tokens in, 2.4K out) costs from $0.0027 on GPT-OSS 20B (Groq) to $0.3829 on Claude Fable 5.1.

List prices checked2026-10-16Standard tierAs JSON By maker

Token cost calculator

Set a review workload; the table below prices it on every model. The page's URL keeps your inputs, so a link shares them.

Model

Input tokens per call

Diff, surrounding code, rules and instructions

Output tokens per call

Findings, reasoning included

Input read from cache (%)

Shared prompt on a panel of one model

Calls per pull request

Reviewers times review rounds

Pull requests per month

For the monthly figure

Sizes are the median small, medium and large reviews among 6,683 successful CodePanel reviewer runs on 958 pull requests, 2026-10-08 to 2026-10-17. Picking one also sets output tokens to what a reviewer wrote at that size.

On Claude Sonnet 5.5, at list price:

Per call: $0.0766
Per pull request: $0.3063
Per month: $30.6336

Every model

Click any column to sort

List prices per 1M tokens and the calculator's workload priced on every model.
				Cache write / 1M
GPT-OSS 20B (Groq)OpenAI · on Groq	$0.075	$0.30	–	–	$0.0108	$1.0763
Gemma 4 26B A4B (Workers AI)Google · on Cloudflare Workers AI	$0.10	$0.30	–	–	$0.0134	$1.3383
Gemini 2.5 Flash-LiteGoogle	$0.10	$0.40	$0.01	–	$0.0144	$1.435
Claude Haiku 5.5AnthropicAbove 100K in: $0.50 / $2.50	$0.10	$0.50	$0.01	$0.125	$0.0153	$1.5317
GPT-6 LunaOpenAIAbove 272K in: $0.20 / $0.75	$0.10	$0.50	$0.01	–	$0.0153	$1.5317
GLM-5.3 Flash (Workers AI)Z.ai · on Cloudflare Workers AI	$0.15	$0.50	–	–	$0.0206	$2.0558
GPT-OSS 120B (Groq)OpenAI · on Groq	$0.15	$0.60	–	–	$0.0215	$2.1525
GPT-5.6 LunaOpenAIAbove 272K in: $0.40 / $1.80	$0.20	$1.20	$0.02	–	$0.0326	$3.2567
GPT-5.4 NanoOpenAI	$0.20	$1.25	$0.02	–	$0.0331	$3.3051
Codestral 25.08Mistral	$0.30	$0.90	–	–	$0.0401	$4.015
Gemini 3.1 Flash-LiteGoogle	$0.25	$1.50	$0.025	–	$0.0407	$4.0709
DeepSeek FlashDeepSeekOff-peak $0.15 / $0.60 · peak 01:00–04:00 and 06:00–10:00 UTC weekdays	$0.30	$1.20	–	–	$0.0431	$4.305
GPT-OSS 120B (Workers AI)OpenAI · on Cloudflare Workers AI	$0.35	$0.75	–	–	$0.0439	$4.3941
MiMo-V2.6-Pro (OpenRouter)Xiaomi · on OpenRouter	$0.435	$0.87	–	–	$0.054	$5.4011
Gemini 2.5 FlashGoogle	$0.30	$2.50	$0.03	–	$0.0556	$5.5618
Gemini 3.5 Flash-LiteGoogle	$0.30	$2.50	$0.03	–	$0.0556	$5.5618
Inkling SmallThinking Machines · on OpenRouter	$0.45	$1.20	–	–	$0.0588	$5.8774
DeepSeek V4 Flash (Workers AI)DeepSeek · on Cloudflare Workers AI	$0.44	$1.32	–	–	$0.0589	$5.8886
Nemotron 3 Super (Workers AI)NVIDIA · on Cloudflare Workers AI	$0.50	$1.50	–	–	$0.0669	$6.6916
Mistral Large 3Mistral · on OpenRouter	$0.50	$1.50	–	–	$0.0669	$6.6916
GLM-5.2Z.ai · on OpenRouter	$0.0341	$7.00	–	–	$0.0713	$7.1251
GLM-4.7Z.ai · on OpenRouter	$0.54	$1.98	–	–	$0.0757	$7.575
Qwen3.8 27B (Workers AI)Qwen · on Cloudflare Workers AI	$0.45	$3.20	–	–	$0.0781	$7.811
Gemini 3.8 FlashGoogleFrom 2027-01-01: $1.50 / $7.50	$0.75	$3.75	$0.075	–	$0.1149	$11.4876
Gemini 3.6 FlashGoogleFrom 2027-01-01: $1.50 / $7.50	$0.75	$3.75	$0.075	–	$0.1149	$11.4876
Gemini 3.7 FlashGoogle · on OpenRouter	$0.75	$3.75	–	–	$0.1149	$11.4876
GPT-5.4 MiniOpenAI	$0.75	$4.50	$0.075	–	$0.1221	$12.2127
Qwen3.8 27B (Groq)Qwen · on Groq	$0.80	$4.00	–	–	$0.1225	$12.2534
Grok Build 0.1xAIAbove 200K in: $2.00 / $4.00	$1.00	$2.00	–	–	$0.1242	$12.4164
Step 5 Preview (OpenRouter)StepFun · on OpenRouter	$1.00	$2.70	–	–	$0.1309	$13.0932
GLM-5Z.ai · on OpenRouter	$1.00	$3.20	–	–	$0.1358	$13.5766
Kimi K2.6Moonshot	$0.95	$4.00	–	–	$0.1383	$13.8259
Kimi K2.7 CodeMoonshot	$0.95	$4.00	–	–	$0.1383	$13.8259
Kimi K2.7 Code (Workers AI)Moonshot · on Cloudflare Workers AI	$0.95	$4.00	–	–	$0.1383	$13.8259
InklingThinking Machines · on OpenRouter	$1.00	$4.05	–	–	$0.144	$14.3983
Claude Haiku 4.5Anthropic	$1.00	$5.00	$0.10	$1.25	$0.1532	$15.3168
Grok 4.3xAIAbove 200K in: $2.50 / $5.00	$1.25	$2.50	–	–	$0.1552	$15.5205
Grok 4.20xAIAbove 200K in: $2.50 / $5.00	$1.25	$2.50	–	–	$0.1552	$15.5205
Muse Spark 1.2Meta · on OpenRouter	$1.25	$4.25	–	–	$0.1721	$17.2124
Muse Spark 1.1Meta · on OpenRouter	$1.25	$4.25	–	–	$0.1721	$17.2124
DeepSeek V4 ProDeepSeekOff-peak $0.66 / $1.98 · peak 01:00–04:00 and 06:00–10:00 UTC weekdays	$1.32	$3.96	–	–	$0.1767	$17.6658
DeepSeek V4 Pro (Workers AI)DeepSeek · on Cloudflare Workers AI	$1.32	$3.96	–	–	$0.1767	$17.6658
Mistral Large 4Mistral · on OpenRouter	$1.36	$4.18	–	–	$0.183	$18.2978
GLM-5.3 (Workers AI)Z.ai · on Cloudflare Workers AI	$1.40	$4.40	–	–	$0.1893	$18.9298
Qwen3.7 MaxQwen · on OpenRouter	$1.475	$4.425	–	–	$0.1974	$19.7402
Gemini 2.5 ProGoogleAbove 200K in: $2.50 / $15.00	$1.25	$10.00	$0.125	–	$0.2277	$22.7715
GPT-5OpenAI · on OpenRouter	$1.25	$10.00	–	–	$0.2277	$22.7715
Mistral Medium 3.5Mistral	$1.50	$7.50	–	–	$0.2298	$22.9752
Grok 4.7xAIAbove 200K in: $4.00 / $12.00	$2.00	$6.00	–	–	$0.2677	$26.7664
Grok 4.6xAIAbove 200K in: $4.00 / $12.00	$2.00	$6.00	–	–	$0.2677	$26.7664
Grok 4.5xAIAbove 200K in: $4.00 / $12.00	$2.00	$6.00	–	–	$0.2677	$26.7664
Qwen3.8 Max (OpenRouter)Qwen · on OpenRouter	$2.00	$6.00	–	–	$0.2677	$26.7664
Kimi K2.7 Code HighspeedMoonshot	$1.90	$8.00	–	–	$0.2765	$27.6517
Claude Sonnet 5.5Anthropic	$2.00	$10.00	$0.10	$2.50	$0.3063	$30.6336
Claude Sonnet 5Anthropic	$2.00	$10.00	$0.20	$2.50	$0.3063	$30.6336
GPT-6.1 SolOpenAIAbove 272K in: $4.00 / $15.00	$2.00	$10.00	$0.10	–	$0.3063	$30.6336
GPT-6 SolOpenAIAbove 272K in: $4.00 / $15.00	$2.00	$10.00	$0.20	–	$0.3063	$30.6336
Gemini 4 ArgonGoogleFrom 2026-12-01: $4.00 / $20.00	$2.00	$10.00	$0.10	–	$0.3063	$30.6336
GPT-5.3 CodexOpenAI	$1.75	$14.00	$0.175	–	$0.3188	$31.8801
GPT-5.2OpenAI · on OpenRouter	$1.75	$14.00	–	–	$0.3188	$31.8801
GPT-5.6 TerraOpenAIAbove 272K in: $4.00 / $18.00	$2.00	$12.00	$0.20	–	$0.3257	$32.5672
Gemini 3.1 ProGoogleAbove 200K in: $4.00 / $18.00	$2.00	$12.00	$0.20	–	$0.3257	$32.5672
GPT-5.4OpenAIAbove 272K in: $5.00 / $22.50	$2.50	$15.00	$0.25	–	$0.4071	$40.709
Claude Sonnet 4.6Anthropic	$3.00	$15.00	$0.30	$3.75	$0.4595	$45.9504
Kimi K3Moonshot	$3.00	$15.00	–	–	$0.4595	$45.9504
MiMo-V2.6-Pro UltraSpeed (OpenRouter)Xiaomi · on OpenRouter	$4.35	$8.70	–	–	$0.5401	$54.0113
Claude Opus 5.5Anthropic	$4.00	$20.00	$0.20	$5.00	$0.6127	$61.2672
GPT-5.6 SolOpenAIAbove 272K in: $8.00 / $30.00	$4.00	$20.00	$0.40	–	$0.6127	$61.2672
Claude Opus 5Anthropic	$5.00	$25.00	$0.50	$6.25	$0.7658	$76.584
Claude Opus 4.8Anthropic	$5.00	$25.00	$0.50	$6.25	$0.7658	$76.584
Claude Opus 4.7Anthropic	$5.00	$25.00	$0.50	$6.25	$0.7658	$76.584
Claude Opus 4.6Anthropic	$5.00	$25.00	$0.50	$6.25	$0.7658	$76.584
GPT-5.5OpenAIAbove 272K in: $10.00 / $45.00	$5.00	$30.00	$0.50	–	$0.8142	$81.418
Claude Fable 5.1Anthropic	$10.00	$50.00	$0.25	$12.50	$1.5317	$153.168
Claude Fable 5Anthropic	$10.00	$50.00	$1.00	$12.50	$1.5317	$153.168
GPT-6 AstraOpenAIAbove 272K in: $20.00 / $75.00	$10.00	$50.00	$1.00	–	$1.5317	$153.168
DeepSeek V4 FlashDeepSeek · retiredOff-peak $0.15 / $0.60 · peak 01:00–04:00 and 06:00–10:00 UTC weekdays	$0.30	$1.20	–	–	$0.0431	$4.305
Gemini 3.5 FlashGoogle · retired	$1.50	$9.00	–	–	$0.2443	$24.4254
GPT-5.5 ProOpenAI · Responses API onlyAbove 272K in: $60.00 / $270.00	$30.00	$180.00	–	–	$4.8851	$488.508
GPT-5.4 ProOpenAI · Responses API onlyAbove 272K in: $60.00 / $270.00	$30.00	$180.00	–	–	$4.8851	$488.508

A dash under cache rates: none in CodePanel's price table, so that input counts at the full input rate. The host may still list one.

What drives the cost of a review

Worked on Claude Sonnet 5.5, the calculator's model.

Pull request size and context: Input is the diff plus the code around it, the repository's rules and the instructions, and it is usually most of the bill. On Claude Sonnet 5.5, one reviewer reading a 12.4K-token pull request costs $0.0405; a 60.9K-token one costs $0.1547.
Reasoning effort: Thinking tokens bill at the output rate, so a higher effort setting costs more on every call. 10K tokens of thinking add $0.10 to a call on Claude Sonnet 5.5, on top of the $0.0766 a typical call costs.
Prompt caching: Every reviewer on a round reads the same pull request, so CodePanel puts that shared material first and has reviewers on the same model and effort share one cache entry: the first writes it, and the rest read it. On Claude Sonnet 5.5, a 26.2K-token prompt costs $0.0524 uncached and $0.0026 read from the cache.
A panel or one model: Each reviewer is its own call. 4 reviewers on Claude Sonnet 5.5 cost $0.3063 a round for typical calls, plus the orchestrator that merges their findings, against $0.0766 for one. Every push can start another round, so a round limit and incremental review after the first round matter as much as the model.

A full worked example, from one round to a team's month: What AI code review costs per pull request.

Questions

What is the cheapest LLM API?

Per review call (26.2K tokens in, 2.4K out), GPT-OSS 20B (Groq) is the cheapest model CodePanel offers, at $0.0027. On a maker's own API the cheapest is Gemini 2.5 Flash-Lite, at $0.0036 a call. The lowest input rate is $0.0341 per 1M tokens, on GLM-5.2. The cheapest model is not the cheapest review: one that misses bugs or raises false alarms costs a second round, so check benchmark results before choosing on price.

Prices as JSON

The same data at /models/pricing.json, for agents and scripts. Every field is always present, and null means it does not apply.

Fields of the pricing JSON.
checked	When the prices were last checked; each entry in sources has its own date.
priceInstant	The instant every rate is read at, so peak hours and scheduled changes resolve the same way each time.
models[].slug	Stable id of the model's page; url is that page.
models[].modelId	The id sent to the host's API.
models[].maker, host	Who made the model, and who serves and prices it. They differ for a model on another company's platform.
models[].input, output	USD per 1M tokens at the standard tier, for a prompt below any long-context threshold.
models[].cacheRead, cacheWrite	USD per 1M cached input tokens read, and written (Anthropic). null means CodePanel's table has no cache rate and bills those tokens as ordinary input, not that the host has none.
models[].longContext	Rates for the whole request once the prompt passes aboveInputTokens, or null.
models[].offPeak	Rates outside peakHoursUtc on weekdays and all weekend, or null.
models[].reprice	A scheduled change: the rates from date, or null.
models[].reviewCallUsd	One typical review call: reviewCall.inputTokens in and reviewCall.outputTokens out.
models[].prInputUsd	Input cost of reading a pull request of each prSizes size.