Model catalogue

41 models from 9 vendors and the open-weight fleet. Prices per 1,000,000 tokens. Rows sorted by weekly token volume.

codingreasoningagents & toolstranslationRP & personaslong contextmultimodalopen weightsself-hostEU residency

ModelContextPrompt, $/MCompletion, $/MQuantizationTok/sTokens this week
gemini-3-flashGooglemultimodal1M ctx1M$0.10$0.40fp824083.6B
qwen-flash-next-3.8Alibabaopen weightsruns locally256k$0.06$0.24fp8 / int416871.3B
gpt-5-miniOpenAIvisiontools256k$0.25$2.00fp814561.4B
deepseek-flash-4.1DeepSeekopen weightsruns locallytools164k$0.08$0.32fp8 / int413258.9B
gpt-5OpenAIvisionreasoning256k$1.25$10.00fp86248.2B
gemini-3-proGooglemultimodal1M ctx1M$1.25$10.00fp811044.8B
qwen3.8-27bAlibabaopen weights24 GB GPUtools128k$0.10$0.40q4 / int89642.7B
claude-sonnet-4.5Anthropiclong taskstools200k$3.00$15.00—7839.5B
deepseek-r2DeepSeekreasoningMoE164k$0.27$1.10fp84535.2B
kimi-k2.5Moonshotopen weightsagentic256k$0.60$2.50fp85824.8B
gpt-5-codexOpenAIcodeagentic400k$1.25$10.00fp85418.7B
llama-4-70bMeta (EU PoP)open weightsEU1M$0.60$1.80fp88815.9B
claude-opus-4.5Anthropictop qualitytools200k$15.00$75.00—4114.3B
glm-4.6Zhipuopen weightsagentic200k$0.30$1.20fp87112.6B
grok-4.1xAIrealtime search512k$2.00$8.00fp86611.2B
devstral-3-coderMistralopen weightscodeлокально256k$0.80$2.40fp8 / int4929.8B
mistral-large-3MistralEUtools128k$2.00$6.00fp8638.4B
llama-4-405bMetaopen weights1M ctx1M$2.50$8.00fp8346.7B
mistral-small-4Mistralopen weightsfast128k$0.20$0.60fp81875.9B
qwen3.8-vl-32bAlibabaopen weightsvisionлокально256k$0.15$0.60fp8 / int4845.2B

Quantization shows the weight format for open-weight models; vendor models do not disclose the exact format (—). Tok/s is the 24-hour median generation speed. Hidden models (21) are available on request for Scale tier.

Where the traffic goes

Coding & code agents34%
Support & business agents21%
Translation & localization14%
Document summarization12%
RP & personas9%
Search & RAG6%
Other4%

How to read the catalogue

  • Open weights — published weights: same API through our fleet, same API on your GPU. qwen-flash-next-3.8, qwen3.8-27b, deepseek-flash-4.1, devstral-3-coder and qwen3.8-vl-32b are mirrored with stable checksums.
  • Runs locally — the model fits 24 GB VRAM at 4-bit quantization; run instructions in the docs.
  • Route modifiers — the :fast, :cheap, :eu and :exact suffixes pick a strategy without changing keys: deepseek-flash-4.1:fast.
  • BYOK — bring your own vendor key on Starter and up: the gateway fee is waived.

Try a model in the chat Pricing