Model Routing

DIT AI says its OpenAI-compatible API routes text, image, and video calls among more than 50 models and 160 providers. The company claims typical costs 30–70% below official API pricing; that savings figure is vendor-reported.

Anthropic’s Boris Cherny said Claude Code users are not banned for using other model harnesses through proxies. He pointed to a classifier issue, unblocked one escalated account, and said false positives are being reduced.

Illscience found Kimi K3 strong on prose reasoning but costly per task because it used more tokens than expected. LLMJunky's Rocket League-style test said the model made good UI but lacked Fable-level polish.

Steipete said he moved the Clawsweeper GitHub review bot to 5.6 Terra high and saw about 40% faster reviews at lower cost with little quality change. A follow-up framed the result as a warning that general model benchmarks may not predict issue-review performance.

Min Choi's workflow assigned Grok 4.5 to research and debugging, Fable 5 to planning and frontend, and GPT-5.6 Sol to complex coding. Shann Holmberg added tactics for effort levels and usage limits.

Kitze described using a $60 Hetzner server, four $200 Codex accounts behind codex-lb, self-hosted Paperclip workspaces, and a GPT-5.5 manager to run about 30 isolated tasks. The setup routes parallel coding work across isolated workspaces and paid Codex accounts.

Anthropic staff said Fable 5 is returning, but classifier-triggered requests can fall back to Opus 4.8 and be billed at Opus rates. That matters for vibe-coders because access, cost, and task routing now depend on whether a request trips the safeguard system.

Runway said its API now includes Seedance 2.0, GPT Image 2, HappyHorse 1.0, Nano Banana Pro, and Magnific Precision Upscaler V2. Try the new cross-model pipeline if you need image-to-video, upscaling, or mixed asset production inside Runway.

GlobalGPT surfaced Seedance 2.0, Wan 2.7, and other video models inside one workspace without invite codes or regional gating in creator tests. The access shift helps rapid model comparison, but today's details come from a single walkthrough thread.

Posts reported an unannounced model called Elephant Alpha appeared on OpenRouter with 100B parameters, 256K context, function calling, and free tokens. No model card or lab has confirmed authorship, so the rankings and efficiency claims remain unattributed.