Modèles frontière, 30–70 % sous le tarif officiel.

Anthropic's most capable model — frontier reasoning, long-horizon agents, 1M context.

Frontier reasoning and long-horizon agents at 1M context — the model Fable 5.1 succeeds.

Near-flagship Opus reasoning at half the price of Fable 5 — vision, agentic coding, 200K context.

Opus 5 tuned for latency — same frontier reasoning, faster output, 200K context.

Anthropic Opus 4.8 — stronger agentic coding and honesty, 1M context.

Anthropic Opus 4.7 — flagship reasoning, vision, 200K context.

Anthropic Opus 4.6 — deep reasoning, exceptional agentic ability.

Near-Opus coding and agentic quality at Sonnet cost — 1M context.

Balanced speed/quality — the everyday production workhorse, elite coding.

Anthropic Haiku 4.5 — fast and cost-efficient.

Google's newest Flash — stronger coding and agentic execution at half the 3.6 official rate, 1M context.

Gemini 3.6 Flash — thinking-by-default at Flash latency, 1M context, native audio input.

Gemini 3.1 Pro — Google's flagship for coding, agents, and cross-modal analysis.

Previous-gen Gemini Flash — extreme value.

OpenAI's newest flagship — top-tier reasoning and agentic coding, 1.05M context.

GPT-5.6 mid tier — the everyday workhorse of the 5.6 family, 1.05M context.

GPT-5.6 fast tier — high-volume, low-latency execution at mini-class cost.

OpenAI GPT-5.4 — strong general reasoning, 1M context.

OpenAI GPT-5.5 — flagship reasoning, 1M context.

OpenAI GPT-5.4 Mini — fast, cost-efficient reasoning.

OpenAI GPT-5.3 Codex — coding-specialized.

OpenAI GPT-5.5 Pro — deep-horizon enterprise reasoning, 1.1M context.

Gemini's native image generation — strong character consistency.

Newer Nano Banana — improved fidelity and prompt adherence.

Gemini 3.1 Flash-Lite Image — ~4s generations for fast, high-volume iteration.

Google's premium Nano Banana — top-tier fidelity and instruction following.

OpenAI's latest native image model — strong prompt understanding.

Image-to-image editor — multi-turn edits with consistency.

GPT-Image-2 image-to-image — edit and remix existing images.

Google Veo 3.1 Fast — native audio, multi-image reference control, extended clips.

Google Veo 3.1 Quality tier — best fidelity, slower + pricier than Fast.

Google Veo 3.1 Lite — the budget tier for storyboarding and previews.

ByteDance Seedance 2 — text-to-video, native audio, references.

Faster, cheaper Seedance 2 — good for previews and high-volume.

ByteDance Seedance 2.5 — up to 30s clips, multimodal references, precise editing.

Cheapest Seedance 2 — 480p/720p only, built for high-volume drafts and previews.

Alibaba Wan 2.7 — strong cinematic text-to-video.

Best-in-class AI music generation — vocals, lyrics, style blending.

Newest Suno — higher fidelity and longer generations.