OpenRouter API Pricing Update: 17 Free Models, New Gemini 4 Argon, DeepSeek Output Price Surge

Seventeen models are free on OpenRouter, led by Google Lyria 3 Pro Preview and Lyria 3 Clip Preview at 1,048,576 context tokens, alongside Thinking Machines Inkling Small and Inkling at the same context length. NVIDIA provides four free Nemotron variants ranging from 128,000 to 1,000,000 context tokens. Poolside offers Laguna S 2.1 and Laguna XS 2.1 at 262,144 tokens, while Google Gemma 4 26B A4B and Gemma 4 31B sit at 262,144 tokens. Cohere North Mini Code provides 256,000 tokens, and the Free Models Router from OpenRouter offers 200,000 tokens. LiquidAI LFM2.5-2.6B rounds out the list at 65,536 tokens.

Seven models joined the catalog. Google Gemini 4 Argon enters at $2.00 input and $10.00 output per million tokens with a 1,048,576 token context, while its batch variant halves those prices to $1.00 and $5.00. LiquidAI LFM2.5-24B-A2B prices at $0.05 input and $0.70 output with 128,000 tokens and open weights. ScaleDown Extract and Summarize both charge $0.04 input with zero output cost at 1,000,000 tokens. Pareto 26.9 and Pareto Latest from unbiased provider each cost $2.50 input and $7.50 output at 262,144 tokens.

Three models were removed: Apodex 1.1 Mini free, inclusionAI Ling 3.0 Flash Fin, and inclusionAI Ling 3.0 Flash Sante. One model stopped being free: inclusionAI Ling 3.1 Flash now charges $0.075 input and $0.225 output per million tokens at 262,144 context. No models became free since the previous day.

DeepSeek V4 Pro 0423 input rose from $0.41 to $0.96 per million tokens and output from $0.82 to $1.91. DeepSeek Pro Latest input fell slightly to $0.225 while output surged from $0.75 to $8.00. DeepSeek V4 Flash 0423 input jumped from $0.0057 to $0.091 with output dropping from $1.28 to $0.182. MoonshotAI Kimi Latest output doubled from $10.00 to $24.00 per million tokens. Meta Llama 3.3 70B Instruct input increased from $0.22 to $0.59 and output from $0.50 to $0.79. Qwen Qwen3.5-35B-A3B input rose from $0.08 to $0.15 and output from $0.75 to $1.00. Z.ai GLM 5.3 Flash and GLM Flash Latest both moved input from $0.0232 to $0.036 while output fell from $1.20 to $0.90. Z.ai GLM Latest input edged up to $0.063 from $0.055 and output dropped to $4.40 from $5.00. DeepSeek V4 Flash Latest input increased to $0.0144 from $0.0128 with output falling to $0.576 from $0.768. inclusionAI Ling 3.0 Flash VL input rose from $0.021 to $0.06 and output from $0.0616 to $0.18. DeepSeek Flash Latest input moved to $0.0405 from $0.0392 and output dropped to $0.765 from $1.188.

DeepSeek V4 Flash Latest leads at $0.0144 input and $0.576 output per million tokens with 1,048,576 context. DeepSeek V4 Flash 0731 follows at $0.016 input and $0.64 output, also at 1,048,576 tokens with open weights. IBM Granite 4.0 Micro charges $0.017 input and $0.112 output at 131,000 tokens with open weights. OpenAI gpt-oss-20b sits at $0.018 input and $0.09 output with 131,072 tokens and open weights. inclusionAI Ling 3.0 Flash prices at $0.021 input and $0.063 output at 262,144 tokens with open weights. OpenAI gpt-oss-20b batch is $0.024 input and $0.112 output. Nex AGI Nex-N2.5-Mini costs $0.025 input and $0.10 output at 262,144 tokens with open weights. OpenAI GPT-5 Nano batch matches $0.025 input with $0.20 output at 400,000 tokens.