Weaving cosmic threads...
Weaving cosmic threads...
Evaluate models by what truly matters for fiction: 1M-token canon retention, poetic cadence, hard magic causality, polyphonic character voices, and anti-slop resistance.
We tested dozens of frontier and open-weight models against 50,000+ words of high fantasy lore, magic system design, and novel chapters. These five represent the absolute peak.
Anthropic
The High Fantasy Novelist & Supreme Prose Craftsman
Run at temperature 0.65 with explicit sensory focus prompts. Best-in-class for chapter manuscripts, lyrical scene descriptions, and razor-sharp dialogue with emotional subtext.
Anthropic
The Arch-Sage of Epic Sagas & Genesis Cosmologies
Reserve for the core spine of your universe: creation myths, pantheon origin scriptures, epic climax chapters, and world-shattering magical confrontations.
The 1M-Token Living Lore Vault & Canon Sentinel
The ultimate lore brain. Drop in your full universe codex, timeline, and 50+ character sheets as system context, then query canon intersections and timeline paradoxes with 99% accuracy.
DeepSeek
The Grand Enchanter of Hard Magic & World Causality
Ask it to audit your magic system for loopholes, exploit balance issues, or simulate geopolitical fallout when an ancient kingdom collapses. Its step-by-step reasoning is extraordinary.
Alibaba
The Sovereign Free Worldbuilder & Bestiary Scribe
The #1 choice for creators building large universes on zero budget. Use for generating hundreds of factions, settlements, creature bestiaries, and flora/fauna encyclopedias.
Anthropic
The High Fantasy Novelist & Supreme Prose Craftsman
Run at temperature 0.65 with explicit sensory focus prompts. Best-in-class for chapter manuscripts, lyrical scene descriptions, and razor-sharp dialogue with emotional subtext.
Anthropic
The Arch-Sage of Epic Sagas & Genesis Cosmologies
Reserve for the core spine of your universe: creation myths, pantheon origin scriptures, epic climax chapters, and world-shattering magical confrontations.
The 1M-Token Living Lore Vault & Canon Sentinel
The ultimate lore brain. Drop in your full universe codex, timeline, and 50+ character sheets as system context, then query canon intersections and timeline paradoxes with 99% accuracy.
DeepSeek
The Grand Enchanter of Hard Magic & World Causality
Ask it to audit your magic system for loopholes, exploit balance issues, or simulate geopolitical fallout when an ancient kingdom collapses. Its step-by-step reasoning is extraordinary.
Mistral AI
The Continental Bard of Polyphonic Dialogue & Court Intrigue
Ideal for writing noble house council arguments, witty bardic banter, espionage debriefs, and subtle power struggles where characters speak in riddles and double meanings.
Alibaba
The Sovereign Free Worldbuilder & Bestiary Scribe
The #1 choice for creators building large universes on zero budget. Use for generating hundreds of factions, settlements, creature bestiaries, and flora/fauna encyclopedias.
OpenAI
Multimodal World Visualizer & Scene Director
Translating textual lore into visual art prompts for image engines (FLUX/Midjourney), or examining visual artifacts and synthesizing lore descriptions.
MiniMax
The Battle Choreographer & Godbeast Clash Director
Drafting epic sword-fights, aerial dragon duels, elemental spell exchanges, and castle sieges. Paces action beats and physical impact with visceral clarity.
DeepSeek
The High-Efficiency World Compendium Scribe
Batch generation of massive world databases: item inventories, potion recipes, plant encyclopedias, minor faction lists, and historical battle tables on a modest budget.
Rapid Multimodal World Explorer & Map Inquisitor
Upload high-resolution hand-drawn fantasy maps and have it analyze geographical realism, suggest mountain pass trade routes, or design climate zones in seconds.
Meta
Sovereign Self-Hosted World Archive & Dark Fantasy Scribe
When you cannot send proprietary book IP to cloud APIs. Excellent for dark fantasy, eldritch horror, and morally grey fiction where proprietary cloud filters cause false refusals.
Moonshot
The Chronology Sentinel & Historical Fact-Checker
Audit your historical timeline for anachronisms, calculate travel times across kingdom maps, and ensure multi-generational characters have logically consistent ages.
Zhipu
The Wandering Folklorist & Ballad Composer
Writing folk songs sung by sailors in port taverns, ancient proverbs, nursery rhymes warning of forest monsters, and regional harvest festival rituals.
Xiaomi
Free 1M Long-Context Saga Outliner
Plotting multi-volume fantasy trilogies where foreshadowing planted in Book 1 must pay off in Book 3.
NVIDIA
The World Demographics & Trade Route Modeler
Designing kingdom economies: tax structures, trade guild monopolies, ore distribution across mountain ranges, and crop yields across river basins.
Zhipu
The Cosmological Thinker & High Priest of Lore
Writing monastic scrolls, theological schisms, and debates on the true nature of the Primordial Duality between Light and Dark.
Anthropic
Rapid World Indexer & Minor NPC Generator
Roll up 20 tavern patrons, city guards, shopkeeper quirks, minor curses, or flora descriptions in 3 seconds.
OpenAI
Lore Metadata Tagger & Glossary Classifier
Automated tagging of existing worldbuilding files, compiling alphabetical indexes, and checking glossary definitions.
OpenAI
Instant In-Editor Fantasy Term & Name Generator
Keep running in the background of your writing app for instant tooltip definitions, synonym generation, and random fantasy name rollups.
inclusionai
inclusionai · not yet reviewed
Ling 3.1 Flash is a hybrid reasoning mixture-of-experts model from inclusionAI, with 25B active parameters out of 560B total.
apodex
apodex · not yet reviewed
Apodex 1.1 Mini is a reasoning-first model from Apodex, built for complex, long-horizon research and forecasting tasks. It works directly with files, data, code, and tools to produce verifiable results,...
unbiased
unbiased · not yet reviewed
Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks. This is a preview of the...
openai
openai · not yet reviewed
GPT-6.1 Sol Pro is the same underlying model as [GPT-6.1 Sol](https://openrouter.ai/openai/gpt-6.1-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...
openai
openai · not yet reviewed
GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional...
anthropic
anthropic · not yet reviewed
Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing...
anthropic
anthropic · not yet reviewed
Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing...
typesafe
typesafe · not yet reviewed
Jev Router picks the best model and reasoning effort for each request, balancing quality, speed, and cost. It runs on [Jev](https://openrouter.ai/~typesafe/jev-latest), TypeSafe's first System One model, and adapts as your...
perceptron
perceptron · not yet reviewed
Perceptron Mk1.5 is Perceptron's embodied reasoning model for physical agents. It accepts text, image, video, and audio input, and answers with text plus optional structured annotations: points, boxes, polygons, tracks,...
fireworks
fireworks · not yet reviewed
Ember-1 is a specialized reasoning model from Fireworks Research, built on [Kimi K3](https://openrouter.ai/moonshotai/kimi-k3). It is designed to make every token go further: it produces shorter reasoning traces, using roughly 40%...
z-ai
z-ai · not yet reviewed
GLM-5.3-Prime is the high-speed variant of Z.ai's GLM-5.3, inheriting its full capabilities while delivering 1.5–2× the output throughput through inference acceleration. It supports text input and output with a 1M-token...
qwen
qwen · not yet reviewed
Qwen3.8 Max Prime is a higher-throughput variant of Qwen3.8 Max from Alibaba's Qwen team, served as a separate SKU at a higher price point. It accepts text, image, and video...
aion-labs
aion-labs · not yet reviewed
Aion 3.5 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It is the smaller, lower-cost sibling of Aion 3.5 and uses...
aion-labs
aion-labs · not yet reviewed
Aion 3.5 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each...
upstage
upstage · not yet reviewed
Solar Mini 4 is Upstage's compact, cost-efficient language model, a 35B-parameter mixture-of-experts with 3B active parameters and a 524K context window. It is built for agentic use cases where response...
cohere
cohere · not yet reviewed
Command A+ is Cohere's flagship model for enterprise agentic workflows. It accepts text and image inputs with a 192K context window, supports native tool calling with strict tool schemas, structured...
openai
openai · not yet reviewed
GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...
openai
openai · not yet reviewed
GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...
openai
openai · not yet reviewed
GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...
openai
openai · not yet reviewed
GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...
openai
openai · not yet reviewed
GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...
openai
openai · not yet reviewed
GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...
openai
openai · not yet reviewed
GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional...
openai
openai · not yet reviewed
GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional...
anthropic
anthropic · not yet reviewed
Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...
anthropic
anthropic · not yet reviewed
Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...
nvidia
nvidia · not yet reviewed
Switchyard is an open-source model router that switches between multiple models to optimize the cost of requests. By default it will use OpenRouter market data to select the most popular...
xiaomi
xiaomi · not yet reviewed
MiMo-V2.6-Pro-UltraSpeed is the fast speed edition of Xiaomi's flagship foundation model, MiMo-V2.6-Pro. Built from the same 1T MiMo-V2.6-Pro checkpoint, it matches the original model in quality while delivering roughly 10x...
xiaomi
xiaomi · not yet reviewed
MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi. Built on a Mixture-of-Experts architecture with 309B total parameters and 15B activated per token, it employs a hybrid attention mechanism for...
Analyze models head-to-head on high fantasy writing metrics: lore memory, prose lyricism, hard magic logic, anti-slop resistance, and full novel generation costs.
The High Fantasy Novelist & Supreme Prose Craftsman
Run at temperature 0.65 with explicit sensory focus prompts. Best-in-class for chapter manuscripts, lyrical scene descriptions, and razor-sharp dialogue with emotional subtext.
The 1M-Token Living Lore Vault & Canon Sentinel
The ultimate lore brain. Drop in your full universe codex, timeline, and 50+ character sheets as system context, then query canon intersections and timeline paradoxes with 99% accuracy.
The Grand Enchanter of Hard Magic & World Causality
Ask it to audit your magic system for loopholes, exploit balance issues, or simulate geopolitical fallout when an ancient kingdom collapses. Its step-by-step reasoning is extraordinary.
Ranked by composite WorldCraft Index, evaluating prose lyricism, 1M canon memory, hard magic logic, anti-slop grade, and token pricing.
| # | Fav | Model & Provider | Gate | WorldCraft | Prose | Lore 1M | Magic | Anti-Slop | Context | In / Out |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 🟤 Claude Sonnet 3.7 / 4🏆 The High Fantasy Novelist & Supreme Prose Craftsman | Voice | 98 | 99% | 96% | 95% | S | 200K | $3.00 / $15.00 | |
| 2 | 🟤 Claude Opus 4 The Arch-Sage of Epic Sagas & Genesis Cosmologies | Crown | 97 | 98% | 97% | 96% | S | 200K | $15.00 / $75.00 | |
| 3 | 🔵 Gemini 2.0 Pro The 1M-Token Living Lore Vault & Canon Sentinel | Starweave | 95 | 90% | 99% | 92% | A | 1M | $1.25 / $5.00 | |
| 4 | 🐋 DeepSeek R1 The Grand Enchanter of Hard Magic & World Causality | Foundation | 94 | 86% | 91% | 99% | A | 128K | $0.55 / $2.19 | |
| 5 | 🦅 Mistral Large 2 The Continental Bard of Polyphonic Dialogue & Court Intrigue | Voice | 94 | 95% | 90% | 90% | S | 128K | $2.00 / $6.00 | |
| 6 | 🟠 Qwen 3.6 PlusFREE The Sovereign Free Worldbuilder & Bestiary Scribe | Unity | 93 | 88% | 98% | 91% | A | 1M | Free | |
| 7 | 🟢 GPT-4o Multimodal World Visualizer & Scene Director | Sight | 92 | 89% | 88% | 89% | B | 128K | $2.50 / $10.00 | |
| 8 | 🟣 MiniMax M2.5FREE The Battle Choreographer & Godbeast Clash Director | Fire | 91 | 89% | 88% | 94% | A | 200K | Free | |
| 9 | 🐋 DeepSeek V3 The High-Efficiency World Compendium Scribe | Foundation | 91 | 88% | 85% | 92% | A | 64K | $0.27 / $1.10 | |
| 10 | 🔵 Gemini 2.0 Flash Rapid Multimodal World Explorer & Map Inquisitor | Sight | 90 | 85% | 97% | 86% | B | 1M | $0.07 / $0.30 | |
| 11 | 🦙 Llama 4 Maverick (3.3 70B)FREE Sovereign Self-Hosted World Archive & Dark Fantasy Scribe | Flow | 89 | 86% | 95% | 88% | B | 1M | Free | |
| 12 | 🌙 Kimi K2.5FREE The Chronology Sentinel & Historical Fact-Checker | Starweave | 88 | 85% | 92% | 87% | B | 260K | Free | |
| 13 | 🟡 GLM 4.7FREE The Wandering Folklorist & Ballad Composer | Flow | 87 | 86% | 88% | 85% | B | 200K | Free | |
| 14 | 🔶 MiMo V2 ProFREE Free 1M Long-Context Saga Outliner | Starweave | 87 | 85% | 94% | 86% | B | 1M | Free | |
| 15 | 💚 Nemotron 3 SuperFREE The World Demographics & Trade Route Modeler | Foundation | 86 | 82% | 94% | 88% | B | 1M | Free | |
| 16 | 🟡 Big Pickle (GLM-4.6)FREE The Cosmological Thinker & High Priest of Lore | Crown | 86 | 84% | 89% | 89% | B | 200K | Free | |
| 17 | 🟤 Claude Haiku 3.5 Rapid World Indexer & Minor NPC Generator | Unity | 85 | 82% | 84% | 82% | B | 200K | $0.25 / $1.25 | |
| 18 | 🟢 GPT-4o Mini Lore Metadata Tagger & Glossary Classifier | Unity | 82 | 80% | 83% | 80% | C | 128K | $0.15 / $0.60 | |
| 19 | 🟢 GPT-5 NanoFREE Instant In-Editor Fantasy Term & Name Generator | Unity | 79 | 76% | 80% | 77% | C | 128K | Free |
Estimate the exact API cost to draft full fantasy manuscripts with your world bible in context. Transparent wholesale token economics.
Note: Free models run at $0.00 via Zen routing. Paid models use direct wholesale provider API pricing.
How Arcanea routes models to dedicated worldbuilding and fantasy novel writing agents. Each agent embodies one of the Ten Gates with Solfeggio tuning, a primary model, and fallback chain.
Physical Geography, Continental Tectonics, Climate & Settlements
Grounds the world in concrete reality. Coordinates landmasses, mountain aquifers, climate belts, trade routes, and settlement eras so the physical world feels ancient and coherent.
Manuscript Chapters, Lyrical Narrative Prose, Pacing & Anti-Slop
Master of the poetic line, rhythmic cadence, and sensory immersion. Crafts full chapters without generic AI-tell verbal tics ("tapestry", "delve"), delivering authentic literary voice.
1M-Token Universe Bibles, Lineages, Timelines & Canon Sentinel
Ingests entire 500-page world compendiums in a single 1M context prompt. Instantly cross-checks multi-generational timelines, family lineages, and prevents canonical contradictions.
Hard Magic Laws, Solfeggio Frequencies, Energy Costs & Runes
Applies rigorous deductive reasoning to spellcraft. Computes energy conservation, consequences, divine oaths, and spell limits so magic feels earned, consistent, and dangerous.
Emotional Interiority, Tragic Flaws, Moral Conflict & Voice
Examines character wounds, hidden yearnings, relationship tensions, and tragic character arcs. Ensures dialogue carries subtext and distinct emotional weight.
Kinetic Magic Battles, Swordplay, Army Sieges & Climactic Duels
Brings intense kinetic velocity, weapon mechanics, tactical terrain advantage, and adrenaline pacing to magical duels, godbeast confrontations, and kingdom-level clashes.
Visual Scene Setting, Citadel Architecture, Atmospheric Weather
Translates raw world ideas into vivid, camera-ready scenes: light filtering through cathedral stained glass, the scent of ozone before a storm, ancient obsidian architecture.
Folktales, Tavern Ballads, Cultural Superstitions & Idioms
Weaves organic world atmosphere through folk songs, regional superstitions, campfire legends, nursery rhymes, and idiomatic expressions across distinct cultures.
Faction Summits, Peace Treaties, Council Debates & Ensemble Casts
Coordinates multi-faction summits and diplomatic arguments. Balances divergent political agendas so every high-table lord speaks with clear ideological distinctiveness.
Primordial Cosmology, Lumina and Nero Duality, Metaphysics
Explores the highest metaphysical layer of existence: the primordial dance of First Light (Lumina) and the Deep Unknown (Nero), divine covenants, and origin of reality.
Pricing and context data come live from provider APIs. Public leaderboards are cited where they exist. WorldCraft scores are editorial ratings, not a published benchmark run.
Hourly synchronization directly from openrouter.ai/api/v1/models. Powers live pricing, context window lengths, token limits, and free tier status.
Human blind double-blind evaluation from the Chatbot Arena Creative Writing & Storytelling category. Measures subjective prose charm, cadence, and emotional connection.
Instruction Following Evaluation (IFEval) for strict negative constraint execution (e.g. banning clichés) plus RULER benchmark for needle-in-haystack recall across 1M tokens.
Editorial ratings for canon consistency, magic-system logic, and anti-slop prose. Assigned by hand from model use, not produced by an automated test run.
Detailed authoring analysis: sensory strengths, fantasy genre sweet spots, anti-slop resistance, and failure modes.
Run at temperature 0.65 with explicit sensory focus prompts. Best-in-class for chapter manuscripts, lyrical scene descriptions, and razor-sharp dialogue with emotional subtext.
Reserve for the core spine of your universe: creation myths, pantheon origin scriptures, epic climax chapters, and world-shattering magical confrontations.
The ultimate lore brain. Drop in your full universe codex, timeline, and 50+ character sheets as system context, then query canon intersections and timeline paradoxes with 99% accuracy.
Ask it to audit your magic system for loopholes, exploit balance issues, or simulate geopolitical fallout when an ancient kingdom collapses. Its step-by-step reasoning is extraordinary.
Ideal for writing noble house council arguments, witty bardic banter, espionage debriefs, and subtle power struggles where characters speak in riddles and double meanings.
The #1 choice for creators building large universes on zero budget. Use for generating hundreds of factions, settlements, creature bestiaries, and flora/fauna encyclopedias.
Translating textual lore into visual art prompts for image engines (FLUX/Midjourney), or examining visual artifacts and synthesizing lore descriptions.
Drafting epic sword-fights, aerial dragon duels, elemental spell exchanges, and castle sieges. Paces action beats and physical impact with visceral clarity.
Track updates to the creative model arena, free tier status, and Solfeggio gate tuning decisions.
Arcanea Model Arena evolved to full Worldbuilding & High Fantasy Intelligence. Claude Sonnet 3.7 / 4 awarded the Editor’s Choice for supreme prose lyricism (99) and anti-slop mastery. Gemini 2.0 Pro awarded Best Lore Vault for 1M-token context retention (99). DeepSeek R1 recognized as Grand Enchanter for hard magic logic (99). 8 models continue running 100% free on Zen routing, led by Qwen 3.6 Plus (1M context) and MiniMax M2.5 (kinetic combat choreography).
Compare FLUX.2, Midjourney v7, DALL-E 3, and Stable Diffusion 3.5. Evaluated for realm cartography, armor heraldry, and godbeast rendering.
Every model in the Arena is directly available in Arcanea Studio. Free models run on Zen routing. Premium models run via your own sovereign API keys.