Claude Fable 5.1anthropicClaude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback)
Intelligence 53.4$7.63/AA task
The latest AI models.
All in one place.
Switch AI models without another subscription.
Chat, image, and video models.
Together in one place.
Docs. Decks. Designs. Just ask.
Featured release · September 10, 2026
DeepSeek V4.1 Flash
Try DeepSeek’s latest Flash model with OpenCraft’s search and file-creation tools.
Explore DeepSeek V4.1 Flash: pricing, vision and capabilities →Meet thefrontier.
Intelligence
GPT-6 AstraopenaiGPT-6 Astra (max)
Intelligence 52.7$3.26/AA task
Muse Spark 1.3metaMuse Spark 1.3 (max)
Intelligence 48.1$1.60/AA task
Qwen3.8 Max (0902)qwenQwen3.8 Max (0902)
Intelligence 45.4$5.41/AA task
GLM 5.3z-aiGLM-5.3 (max) · Open weights
Intelligence 44.8$2.01/AA task
Images
GPT Image 2openaiGPT Image 2 (high)
Grok Imagine Image 2.0spacexaiGrok Imagine Image 2.0
Elo 1154 ±12$0.06/img
Muse Image 1.0metaMuse Image
Elo 1111 ±10$0.01/img
Seedream 5.0 ProbytedanceSeedream 5.0 Pro
Elo 1078 ±8$0.035/img
Video
Wan v3.0 VideoalibabaWan 3.0 · No audio
Elo 1336 ±13$0.05/s
MiniMax H3minimaxMiniMax H3 · No audio
Elo 1302 ±11$0.08/s
Grok Imaginespacexaigrok-imagine-video · No audio
Elo 1235 ±8$0.05/s
Kling v3.0 Text-to-VideoklingaiKling 3.0 1080p (Pro) · No audio
Elo 1230 ±8$0.168/s
Veo 3.1 Lite GenerategoogleVeo 3.1 Lite · No audio
Elo 1216 ±10$0.03/s
Language costs are measured AA benchmark tasks. Image/video prices are serving rates and vary with settings.
Rankings: Artificial Analysis · Updated 21 Sept 2026. Highest reasoning effort with both intelligence and task cost measured. One model per provider; the Intelligence and Efficiency groups each include published weights. Image/video ranks use separate text-to-image and text-to-video (without audio) Elo arenas; ± shows the 95% confidence interval. Tested settings may differ from your generation settings.
More forless.
The efficiency tier. Lower costs for everyday work.
GLM 5.3 Flash
z-ai
AA intelligence 41.8 · Default configuration · Open weights
≈30×benchmark tasks per dollar
$0.253 / AA task
GPT-5.6 Terra
openai
AA intelligence 42.1 · max
≈5×benchmark tasks per dollar
$1.40 / AA task
Compared with Claude Fable 5.1 at $7.63 / AA task. Picks follow the AA intelligence–task cost frontier (index v4.3), above 75% of the leading score and cheaper than the Intelligence group. AA benchmark task costs are not OpenCraft task quotes.
Why pay frontier prices
for every step?
Illustrative long-session token budget
Model cost for the same token volume
OpenCraft using GPT‑5.6 Luna. 360+ models on OpenCraft, with new ones added daily. Hypothetical uncached token-only calculation — 3M tokens in, 130K out — at August 24, 2026 list prices. It is not a measured deck run and excludes tools and OpenCraft's credit purchase fee.
No lock‑in.
Use the best models from any provider. The day they ship.