Compare AI models

102 curated head-to-head comparisons with live pricing, context windows and capabilities from the registry Zeplik routes and bills against. Every model in a comparison is available to chat with on Zeplik.

Assistant against assistant

• Claude vs ChatGPTLive prices and context windows, plus which to pick for what.
• ChatGPT vs GeminiLive prices and context windows, plus which to pick for what.

Across providers

• Claude Opus 5.5 vs GPT-6 AstraAnthropic vs OpenAIAnthropic's and OpenAI's deep-reasoning flagships, eighteen days apart -- and Astra asks two and a half times Opus 5.5's price for a window only five percent larger.
• Claude Opus 5.5 vs GPT-6 SolAnthropic vs OpenAIBoth shipped on the same day at adjacent tiers: whether Opus 5.5's depth is worth double GPT-6 Sol's price when their context windows are within five percent of each other.
• Claude Opus 5.5 vs Grok 4.7Anthropic vs SpaceXAITwo flagships a day apart: Opus 5.5 doubles Grok 4.7's context and asks roughly four times as much per output token.
• Claude Opus 5.5 vs DeepSeek V4 Pro 0423Anthropic vs DeepSeekAnthropic's premium reasoning tier against the open-weight model that answers for a fraction of the output price, and cannot read an image at all.
• GPT-6 Sol vs Claude Sonnet 5OpenAI vs AnthropicPriced identically at $2 in and $10 out, five percent apart on context: the closest like-for-like comparison in the catalogue, and a genuine toss-up on price alone.
• GPT-6 Sol vs Grok 4.7OpenAI vs SpaceXAITwo September flagships a day apart: Grok 4.7 undercuts GPT-6 Sol by half on output, and GPT-6 Sol answers with twice the context window.
• GPT-6 Luna vs Gemini 3.8 FlashOpenAI vs GoogleThe value tier at its current edge: GPT-6 Luna undercuts Gemini Flash by seven and a half times on both legs, with a marginally larger window.
• GPT-6 Luna vs DeepSeek V4 Flash 0423OpenAI vs DeepSeekThe honest throughput comparison: DeepSeek's flash tier is cheaper still per output token, and it cannot read an image, which GPT-6 Luna can.
• Grok 4.7 vs Claude Sonnet 5SpaceXAI vs AnthropicGrok 4.7 answers for roughly half Sonnet 5's output price; Sonnet 5 answers with twice the context. The trade is unusually clean.
• GPT-6 Astra vs Claude Opus 5OpenAI vs AnthropicThe new top-tier head-to-head: OpenAI's GPT-6 flagship against Anthropic's deepest model, which asks half the price for a comparable million-token window.
• GPT-6 Astra vs Claude Sonnet 5OpenAI vs AnthropicFrontier against workhorse: whether GPT-6 Astra earns five times Sonnet 5's price on the work you actually send it.
• GPT-6 Astra vs Gemini 3.1 Pro PreviewOpenAI vs GoogleOpenAI's GPT-6 flagship against Gemini Pro, which costs a fifth as much per input token and accepts audio and video that Astra cannot read.
• GPT-6 Astra vs Grok 4.6OpenAI vs SpaceXAIThe GPT-6 flagship against xAI's current top model, at roughly eight times the output price and twice the context.
• GPT-6 Astra vs Claude Fable 5.1OpenAI vs AnthropicTwo models priced identically at $10 in and $50 out, aimed at opposite ends of the work: Astra's long-horizon engineering against Fable's literary voice.
• GPT-6 Astra vs DeepSeek V4 Pro 0423OpenAI vs DeepSeekThe most expensive closed flagship against the open-weight model that answers for about a thirtieth of the output price -- and the vision gap that comes with it.
• GPT-5.6 Sol vs Claude Sonnet 5OpenAI vs AnthropicThe two defaults most people actually choose between: OpenAI's mainstream flagship against Anthropic's balanced current generation.
• GPT-5.6 Sol vs Claude Fable 5OpenAI vs AnthropicOpenAI's flagship against Anthropic's writing-and-reasoning specialist from the same release wave.
• GPT-5.6 Sol vs Claude Opus 4.8OpenAI vs AnthropicThe head-to-head at the very top: each lab's premium tier on the hardest work.
• GPT-5.6 Sol vs Claude Opus 5OpenAI vs AnthropicThe new head-to-head at the very top: OpenAI's flagship against Anthropic's million-token Opus 5 premium tier.
• GPT-5.6 Sol vs Gemini 3.1 Pro PreviewOpenAI vs GoogleOpenAI against Google at the capability tier, with Gemini's huge context window as the wildcard.
• GPT-5.6 Sol vs Grok 4.6OpenAI vs SpaceXAIThe GPT flagship against xAI's current top model.
• Claude Sonnet 5 vs Grok 4.6Anthropic vs SpaceXAIAnthropic's balanced flagship against Grok 4.6, at the same tier of the market.
• GPT-5.6 Sol vs Grok 4.5OpenAI vs SpaceXAIThe GPT flagship against xAI's reasoning-first Grok 4.5.
• GPT-5.6 Sol vs DeepSeek V4 Pro 0423OpenAI vs DeepSeekClosed flagship against the open-weight challenger that undercuts it dramatically on price.
• Claude Sonnet 5 vs Gemini 3.1 Pro PreviewAnthropic vs GoogleAnthropic's balanced flagship against Google's long-context capability tier.
• Claude Sonnet 5 vs Grok 4.5Anthropic vs SpaceXAITwo reasoning-forward flagships with very different temperaments.
• Claude Sonnet 5 vs DeepSeek V4 Pro 0423Anthropic vs DeepSeekAnthropic's flagship against the strongest open-weight alternative for a fraction of the cost.
• Claude Opus 4.8 vs Grok 4.5Anthropic vs SpaceXAIAnthropic's deepest model against xAI's flagship on frontier-difficulty work.
• Claude Opus 4.8 vs Gemini 3.1 Pro PreviewAnthropic vs GooglePremium-tier depth against premium-tier context: Opus quality versus Gemini scale.
• Gemini 3.1 Pro Preview vs Grok 4.5Google vs SpaceXAIGoogle's capability tier against xAI's, two labs that iterate fastest on knowledge freshness.
• DeepSeek V4 Pro 0423 vs Grok 4.5DeepSeek vs SpaceXAIOpen-weight value against closed frontier reasoning.
• Gemini 3.1 Pro Preview vs DeepSeek V4 Pro 0423Google vs DeepSeekGoogle's flagship against the open-weight model that matches much of it at open pricing.
• Llama 4 Maverick vs DeepSeek V4 Pro 0423Meta vs DeepSeekThe two most-watched open-weight flagships, Meta scale against DeepSeek reasoning.
• Llama 4 Maverick vs Qwen3.7 MaxMeta vs QwenMeta's open flagship against Alibaba's top Qwen tier.
• DeepSeek V4 Pro 0423 vs Qwen3.7 MaxDeepSeek vs QwenThe two leading Chinese open-model programs at their respective tops.
• DeepSeek V4 Pro 0423 vs GLM 5.2DeepSeek vs Z.aiDeepSeek against Z.ai's GLM, both open-weight lines with strong coding results.
• Qwen3.7 Max vs GLM 5.2Qwen vs Z.aiQwen's capability tier against the GLM 5.2 generation.
• Kimi K2 Thinking vs DeepSeek V4 Pro 0423MoonshotAI vs DeepSeekKimi's deliberate thinking variant against DeepSeek's flagship reasoner.
• MiniMax M3 vs GLM 5.2MiniMax vs Z.aiTwo fast-iterating challengers, MiniMax's M3 against GLM 5.2.
• MiniMax M3 vs DeepSeek V4 Pro 0423MiniMax vs DeepSeekMiniMax's M3 against the open-weight benchmark everyone measures by.
• Llama 4 Maverick vs Mistral Medium 3.5Meta vs MistralMeta's open flagship against Europe's leading lab at its capability tier.
• GPT-5.6 Luna vs Gemini 3.7 FlashOpenAI vs GoogleThe value tier where volume work actually runs: OpenAI's Luna against Gemini 3.7 Flash, at a fraction of either lab's flagship price.
• Gemini 3.7 Flash vs DeepSeek V4 Flash 0423Google vs DeepSeekA closed flash tier against the open-weight one, both priced for throughput rather than depth.
• Claude Sonnet 5 vs Gemini 3.8 FlashAnthropic vs GoogleA flagship against a flash tier that costs a fraction of it: the honest question is how much of your work actually needs the flagship.
• Claude Sonnet 5 vs Gemini 3.7 FlashAnthropic vs GoogleA flagship against a flash tier that costs a fraction of it: the honest question is how much work actually needs the flagship.
• GPT-5.6 Luna vs Gemini 3.5 FlashOpenAI vs GoogleThe value tier where volume work actually runs: OpenAI's Luna against Gemini Flash.
• GPT-5.6 Luna vs DeepSeek V4 Flash 0423OpenAI vs DeepSeekOpenAI's efficient tier against DeepSeek's open-weight speed variant.
• Gemini 3.5 Flash vs DeepSeek V4 Flash 0423Google vs DeepSeekTwo flash tiers built for throughput, closed against open.
• GPT-5.6 Luna vs Mistral Small 4OpenAI vs MistralCompact price-performance from OpenAI and Mistral.
• Gemini 3.5 Flash vs Mistral Small 4Google vs MistralGemini's speed tier against Mistral's efficient Small.
• GPT-5.6 Luna vs Claude Sonnet 5OpenAI vs AnthropicOpenAI's value tier against Anthropic's flagship: when is the cheaper model enough?
• Llama 4 Scout vs Gemini 3.5 FlashMeta vs GoogleMeta's efficient long-context Scout against Gemini Flash.
• Grok Build 0.1 vs Kimi K2.7 CodeSpaceXAI vs MoonshotAITwo dedicated coding models, xAI's Grok Build against Moonshot's Kimi Code.
• Grok Build 0.1 vs Claude Sonnet 5SpaceXAI vs AnthropicA coding specialist against the general flagship engineers most often reach for.
• Kimi K2.7 Code vs Codestral 2508MoonshotAI vs MistralOpen-weight code specialists from Moonshot and Mistral.
• GPT-5.6 Sol vs Grok Build 0.1OpenAI vs SpaceXAIOpenAI's generalist flagship against a purpose-built coding model.
• Gemini 4 Argon vs GPT-6 SolGoogle vs OpenAIGoogle's newest flagship against OpenAI's mid tier, priced identically at $2 in and $10 out: the closest cross-lab like-for-like in the catalogue, and a decision price cannot settle.
• DeepSeek V4.1 Flash vs GPT-6 LunaDeepSeek vs OpenAIDeepSeek's newest flash tier against OpenAI's value tier: the two cheapest serious options here, and neither one can read an image.
• Kimi K3.1 Pro vs GPT-6 SolMoonshotAI vs OpenAIMoonshot's newest against GPT-6 Sol, priced identically at $2 in and $10 out: price will not decide this one, so it comes down to which answers your work better.

Within a family

• Claude Opus 5.5 vs Claude Opus 5Anthropic vs AnthropicThe rare upgrade with nothing to weigh: Opus 5.5 is newer than Opus 5 and twenty percent cheaper on both input and output, at the same million-token window.
• Claude Opus 5.5 vs Claude Sonnet 5Anthropic vs AnthropicThe current Anthropic tier decision: Opus 5.5 costs twice Sonnet 5 for an identical million-token window, so the question is entirely about depth per turn.
• Claude Opus 5.5 vs Claude Fable 5.1Anthropic vs AnthropicA reversal worth knowing about: Anthropic's deep reasoning tier now costs less than half its writing specialist, where the deepest model used to be the dearest.
• GPT-6 Sol vs GPT-5.6 SolOpenAI vs OpenAIThe generational swap that also halves the bill: GPT-6 Sol replaces GPT-5.6 Sol at $2 in and $10 out, half what the model it supersedes now asks, on the same 1,050,000-token window.
• GPT-6 Sol vs GPT-6 AstraOpenAI vs OpenAIThe upgrade decision inside the GPT-6 line: Astra costs five times Sol for the same 1,050,000-token window, so the question is which work is hard enough to need it.
• GPT-6 Sol vs GPT-6 LunaOpenAI vs OpenAIGPT-6's capability tier against its value tier, twenty times apart on both input and output price with the same context window on each.
• GPT-6 Sol vs GPT-6 Sol ProOpenAI vs OpenAISol against its Pro serving at an identical price: what more computation per answer actually changes, when the token cost does not.
• GPT-6 Luna vs GPT-5.6 LunaOpenAI vs OpenAIThe cheapest upgrade on the board: GPT-6 Luna halves its predecessor's input price and cuts output by well over half, at the same 1,050,000-token window.
• Grok 4.7 vs Grok 4.6SpaceXAI vs SpaceXAIThe newest Grok against the one it replaces, priced the same on both legs with the same 500,000-token window: a straight generational question.
• Grok 4.7 vs Grok Build 0.1SpaceXAI vs SpaceXAIxAI's newest generalist against its own coding specialist, which is cheaper per token but reads half the context.
• GPT-6 Astra vs GPT-5.6 SolOpenAI vs OpenAIThe upgrade decision almost every OpenAI user is actually making: GPT-6 Astra costs two and a half times GPT-5.6 Sol per token for the same 1,050,000-token window, so the question is which work is hard enough to need it.
• GPT-6 Astra vs GPT-6 Astra ProOpenAI vs OpenAIThe same underlying model at the same price, served with reasoning mode set to pro: what the extra computation per answer actually changes.
• GPT-6 Astra vs GPT-5.6 LunaOpenAI vs OpenAIThe widest price gap inside OpenAI's own catalog -- forty times per output token -- between the GPT-6 flagship and the tier most volume work runs on.
• Gemini 3.8 Flash vs Gemini 3.7 FlashGoogle vs GoogleThe current Gemini Flash against the one it replaces: a straight generational question, with each one's live posted price in the table below.
• GPT-5.6 Sol vs GPT-5.6 LunaOpenAI vs OpenAIOpenAI's capability tier against its value tier: what the extra spend buys.
• GPT-5.6 Sol vs GPT-5.6 TerraOpenAI vs OpenAITwo tiers of the same GPT-5.6 generation, Sol above, Terra below.
• GPT-5.6 Luna vs GPT-5.6 TerraOpenAI vs OpenAIThe two efficiency tiers of GPT-5.6 head to head.
• GPT-5.6 Sol vs GPT-5.6 Sol ProOpenAI vs OpenAISol against its Pro serving: same model family, more computation per answer.
• Claude Sonnet 5 vs Claude Fable 5Anthropic vs AnthropicAnthropic's two current-generation flagships, balanced Sonnet against literary Fable.
• Claude Sonnet 5 vs Claude Opus 4.8Anthropic vs AnthropicSonnet's balance against Opus depth, the classic Anthropic tier decision.
• Claude Opus 5 vs Claude Sonnet 5Anthropic vs AnthropicThe Anthropic tier decision one generation back: whether Opus 5's depth and million-token window justify two and a half times Sonnet's price.
• Claude Opus 5 vs Claude Fable 5Anthropic vs AnthropicAnthropic's deepest new model against its writing specialist: raw capability versus a tuned literary voice.
• Claude Opus 4.8 vs Claude Opus 4.7Anthropic vs AnthropicOne Opus generation against the last: what actually improved.
• Gemini 3.7 Flash vs Gemini 3.6 FlashGoogle vs GoogleOne Gemini Flash generation to the next, three weeks apart and priced identically: whatever you are choosing between here, it is not cost.
• Gemini 3.7 Flash vs Gemini 3.5 FlashGoogle vs GoogleThe upgrade most Gemini Flash users are actually weighing: 3.7 against the 3.5 they are probably still on, for half the input cost and well under half the output.
• Gemini 3.1 Pro Preview vs Gemini 3.7 FlashGoogle vs GoogleGemini Pro capability against Gemini 3.7 Flash speed inside one catalog.
• Gemini 3.1 Pro Preview vs Gemini 3.5 FlashGoogle vs GoogleGemini Pro capability against Gemini Flash speed inside one catalog.
• Gemini 3.5 Flash vs Gemini 3.1 Flash LiteGoogle vs GoogleGemini's two speed tiers, Flash against the even lighter Flash-Lite.
• Grok 4.6 vs Grok 4.5SpaceXAI vs SpaceXAIGrok 4.6 against the release it replaces, at an identical price and an identical 500,000-token window.
• Grok 4.5 vs Grok 4.3SpaceXAI vs SpaceXAIGrok 4.5 against the release it replaced, which was cheaper per token and read twice the context.
• Grok 4.5 vs Grok Build 0.1SpaceXAI vs SpaceXAIxAI's generalist flagship against its own coding specialist.
• DeepSeek V4 Pro 0423 vs DeepSeek V4 Flash 0423DeepSeek vs DeepSeekDeepSeek's capability tier against its throughput tier in the same generation.
• DeepSeek V4 Pro 0423 vs DeepSeek V3.2DeepSeek vs DeepSeekV4 against V3.2: the generational jump in DeepSeek's line.
• Llama 4 Maverick vs Llama 4 ScoutMeta vs MetaLlama 4's two personalities, Maverick capability against Scout efficiency.
• Llama 4 Maverick vs Llama 3.3 70B InstructMeta vs MetaLlama 4 against the 3.3 workhorse a huge install base still runs.
• MiniMax M3 vs MiniMax M2.7MiniMax vs MiniMaxMiniMax's M3 against the best of the M2 line.
• Qwen3.7 Max vs Qwen3.7 PlusQwen vs QwenQwen's Max capability tier against the cost-balanced Plus.
• Kimi K2.7 Code vs Kimi K2 ThinkingMoonshotAI vs MoonshotAIKimi's coding specialist against its extended-thinking variant.
• GLM 5.2 vs GLM 5.1Z.ai vs Z.aiGLM 5.2 against its immediate predecessor, at a third less per token and five times the context window.
• Mistral Medium 3.5 vs Mistral Small 4Mistral vs MistralMistral's capability tier against its efficiency tier.
• GPT-6.1 Terra vs GPT-6 SolOpenAI vs OpenAIOpenAI's newest against the GPT-6 tier it refines: Terra asks half Sol's price per token on the same 1,050,000-token window, which makes this a rare upgrade with no trade to weigh.
• Claude Haiku 5.5 vs Claude Sonnet 5Anthropic vs AnthropicAnthropic's newest tier is also its cheapest: Haiku 5.5 answers for a twentieth of Sonnet 5's price on an identical million-token window, so the only question is how much depth the work needs.
• MiniMax M3.1 Flash vs MiniMax M3MiniMax vs MiniMaxMiniMax's newest flash tier against the M3 it extends: where throughput is bought and what it costs in depth.

Browse all models