v0.1.42 — 273 models, 44 providers

What's new in v0.1.42

New provider

• Perceptron (#44): video+embodied AI vision-language models

New models (267 → 273)

• Qwen3.6 series (4 models): Flash (1M ctx), 35B MoE, 27B, Max Preview
• Qwen3.5 Plus v2: Apr 2026 update (1M ctx, $0.30/$1.80)
• DeepSeek V4 Pro: 1.6T-param MoE flagship ($0.435/$0.87, 1M ctx)
• DeepSeek V4 Flash: 284B-param efficient MoE ($0.112/$0.224, 1M ctx)
• Mistral Medium 3.5: dense 128B ($1.5/$7.5, 262k ctx)
• Moonshot Kimi K2.6: next-gen multimodal ($0.73/$3.49, 262k ctx)
• Perceptron Mk1: video+embodied VLM ($0.15/$1.5, 32k ctx)

Stats

267 → 273 models | 43 → 44 providers