New Text Models
META Llama 3.1 (405B)
Meta’s flagship open-weight model at 405 billion parameters. Competitive with GPT-4 class models on many benchmarks. Available in AI Lab at cost with no markup.
META Llama 3.1 (70B)
The 70B version offers a strong quality-to-cost ratio. Good for teams that want open-weight models without the latency of the 405B.
Perplexity Sonar Large
Perplexity’s online model with real-time web search integration. Useful for queries that require current information.
GPT-4o mini
OpenAI’s lighter GPT-4o variant. Substantially faster and lower-cost — a strong option for high-volume, lower-complexity tasks.
New Image Models
FLUX.1 Pro (Black Forest Labs)
High-quality image generation with strong prompt fidelity. One of the most capable models currently available for realistic imagery.
Leonardo v1 (Phoenix predecessor)
Leonardo’s prior-generation model retained for comparison alongside the newer Phoenix release.
Stable Diffusion Image Ultra
Stability AI’s top-tier image model in the SD lineup.
Amazon Titan v2
Amazon’s image generation model via Bedrock API.
Price Adjustments
Gemini model pricing updated to reflect current Google AI API rates. All prices in AI Lab reflect what we pay providers — no markup.