Gateway catalog changelog

• DeepSeek V4.1 Flash Heretic Thinking arrived, first seen on NanoGPT
• DeepSeek V4.1 Flash Heretic arrived, first seen on NanoGPT
• Phi 4 Multimodal arrived, first seen on NanoGPT
• Phi 4 Mini arrived, first seen on NanoGPT
• weaver arrived, first seen on Experiential Labs
• unslopnemo-12b arrived, first seen on Experiential Labs
• ui-tars-1.5-7b arrived, first seen on Experiential Labs
• sonar-reasoning-pro arrived, first seen on Experiential Labs
• sonar-pro-search arrived, first seen on Experiential Labs
• sonar-pro arrived, first seen on Experiential Labs
• seed-2.0-mini arrived, first seen on Experiential Labs
• seed-2.0-lite arrived, first seen on Experiential Labs

See all 221

• Isoquant added glm-5.3-flash
• Isoquant added deepseek-v4.1-flash-1017-fast
• Isoquant added deepseek-v4.1-flash-1017
• Isoquant added deepseek-v4.1-flash-1017-fast
• NanoGPT added DeepSeek V4.1 Flash Heretic Thinking
• NanoGPT added DeepSeek V4.1 Flash Heretic
• NanoGPT added Phi 4 Multimodal
• NanoGPT added Phi 4 Mini
• NanoGPT added MiniMax M1
• NanoGPT added Mercury Coder Small
• NanoGPT added Mercury 2
• NanoGPT added LongCat 2.0

See all 1298

• Isoquant cut input 70% to $0.030 and output 17.8% to $0.370 on deepseek-v4.1-flash-1017
• NanoGPT raised input 11.4% to $0.078 on GLM 5.2 Thinking
• NanoGPT raised input 11.4% to $0.078 on GLM 5.2
• NanoGPT cut input 52% to $0.120 on Qwen3 Coder Next
• NanoGPT cut input 66.7% to $0.100 and output 80% to $0.100 on Qwen 3 235b A22B 2507 Thinking
• NanoGPT raised input 53.8% to $0.020 and output 88.7% to $0.100 on GPT OSS 20B
• NanoGPT raised input 100% to $4.00 and output 100% to $20.00 on GPT 5.6 Sol Pro
• NanoGPT raised input 100% to $4.00 and output 100% to $20.00 on GPT 5.6 Sol
• NanoGPT raised output 9.8% to $1.80 on Kimi K2.6 Thinking
• NanoGPT raised output 9.8% to $1.80 on Kimi K2.6
• NanoGPT cut input 4.8% to $0.200 and raised output 42.9% to $1.20 on MiniMax M2.7
• NanoGPT cut input 25.9% to $0.200 and raised output 11.1% to $1.20 on MiniMax M2.5

See all 176

• NanoGPT raised Qwen3.8 Flash's context window from 991.8K to 1M
• NanoGPT lowered Qwen3.5 122B A10B Thinking's context window from 262.1K to 131.1K
• NanoGPT lowered Qwen3.5 122B A10B's context window from 262.1K to 131.1K

• Cheaper Inference dropped gpt-5.1
• Cheaper Inference dropped laya-free
• Cheaper Inference dropped claude-haiku-4.5-20251001-r
• LLM Gateway dropped Qwen3 Reranker 8B

• laya-free is no longer served by any gateway we track
• claude-haiku-4.5-20251001-r is no longer served by any gateway we track

• Experiential Labs stopped publishing a price for laguna-xs-2.1
• Experiential Labs stopped publishing a price for laguna-s-2.1

• Phi 4 Multimodal arrived, first seen on NanoGPT
• Phi 4 Mini arrived, first seen on NanoGPT
• weaver arrived, first seen on Experiential Labs
• unslopnemo-12b arrived, first seen on Experiential Labs
• ui-tars-1.5-7b arrived, first seen on Experiential Labs
• sonar-reasoning-pro arrived, first seen on Experiential Labs
• sonar-pro-search arrived, first seen on Experiential Labs
• sonar-pro arrived, first seen on Experiential Labs
• seed-2.0-mini arrived, first seen on Experiential Labs
• seed-2.0-lite arrived, first seen on Experiential Labs
• seed-1.6-flash arrived, first seen on Experiential Labs
• qwen3.5-plus-20260420 arrived, first seen on Experiential Labs

See all 221

• Isoquant added glm-5.3-flash
• Isoquant added deepseek-v4.1-flash-1017
• NanoGPT added Phi 4 Multimodal
• NanoGPT added Phi 4 Mini
• NanoGPT added MiniMax M1
• NanoGPT added Mercury Coder Small
• NanoGPT added Mercury 2
• NanoGPT added LongCat 2.0
• NanoGPT added Gemma 4 12B Instruct
• NanoGPT added Gemini 3 Pro Image
• NanoGPT added Gemini 2.5 Pro Preview 0605
• NanoGPT added Gemini 2.5 Pro Preview 0506

See all 1296

• NanoGPT cut input 18.6% to $0.070 on GLM 5.2 Thinking
• NanoGPT cut input 18.6% to $0.070 on GLM 5.2
• NanoGPT cut input 35% to $0.013 and output 47% to $0.053 on GPT OSS 20B
• NanoGPT cut input 45.8% to $0.149 and output 28.1% to $7.15 on Kimi K3
• NanoGPT cut output 8.9% to $1.64 on Kimi K2.6 Thinking
• NanoGPT cut output 8.9% to $1.64 on Kimi K2.6
• NanoGPT raised input 10% to $0.011 and output 10% to $0.033 on Gemma 4 26B A4B Thinking
• NanoGPT raised input 10% to $0.011 and output 10% to $0.033 on Gemma 4 26B A4B
• NanoGPT raised input 103.2% to $0.045 and cut output 6.9% to $0.179 on DeepSeek V4.1 Flash Thinking
• NanoGPT raised input 103.2% to $0.045 and cut output 6.9% to $0.179 on DeepSeek V4.1 Flash
• NanoGPT cut input 57% to $0.284 and output 57% to $0.851 on DeepSeek V4 Pro 0813 Thinking
• NanoGPT cut input 57% to $0.284 and output 57% to $0.851 on DeepSeek V4 Pro 0813

See all 159

• Cheaper Inference dropped venice-recraft-v4-pro
• Cheaper Inference dropped venice-qwen-image-2
• Cheaper Inference dropped venice-qwen-image
• Cheaper Inference dropped venice-parakeet-asr
• Cheaper Inference dropped venice-hunyuan-image-v3
• Cheaper Inference dropped venice-gpt-image-2
• Cheaper Inference dropped venice-gpt-image-1.5
• Cheaper Inference dropped venice-flux-2-pro
• Cheaper Inference dropped venice-flux-2-max
• Cheaper Inference dropped uncensored-wan-2.2-ultra-fast
• Cheaper Inference dropped uncensored-t2v-turbo
• Cheaper Inference dropped uncensored-t2v-standard

See all 76

• venice-recraft-v4-pro is no longer served by any gateway we track
• venice-qwen-image-2 is no longer served by any gateway we track
• venice-qwen-image is no longer served by any gateway we track
• venice-parakeet-asr is no longer served by any gateway we track
• venice-hunyuan-image-v3 is no longer served by any gateway we track
• venice-gpt-image-2 is no longer served by any gateway we track
• venice-gpt-image-1.5 is no longer served by any gateway we track
• venice-flux-2-pro is no longer served by any gateway we track
• venice-flux-2-max is no longer served by any gateway we track
• uncensored-wan-2.2-ultra-fast is no longer served by any gateway we track
• uncensored-t2v-turbo is no longer served by any gateway we track
• uncensored-t2v-standard is no longer served by any gateway we track

See all 60

• Phi 4 Multimodal arrived, first seen on NanoGPT
• Phi 4 Mini arrived, first seen on NanoGPT
• OrcaCyber Zero 1.5 arrived, first seen on NanoGPT
• weaver arrived, first seen on Experiential Labs
• unslopnemo-12b arrived, first seen on Experiential Labs
• ui-tars-1.5-7b arrived, first seen on Experiential Labs
• sonar-reasoning-pro arrived, first seen on Experiential Labs
• sonar-pro-search arrived, first seen on Experiential Labs
• sonar-pro arrived, first seen on Experiential Labs
• seed-2.0-mini arrived, first seen on Experiential Labs
• seed-2.0-lite arrived, first seen on Experiential Labs
• seed-1.6-flash arrived, first seen on Experiential Labs

See all 231

• Isoquant added glm-5.3-flash
• Isoquant added deepseek-v4.1-flash-1017
• NanoGPT added Phi 4 Multimodal
• NanoGPT added Phi 4 Mini
• NanoGPT added OrcaCyber Zero 1.5
• NanoGPT added MiniMax M1
• NanoGPT added Mercury Coder Small
• NanoGPT added Mercury 2
• NanoGPT added LongCat 2.0
• NanoGPT added Gemma 4 12B Instruct
• NanoGPT added Gemini 3 Pro Image
• NanoGPT added Gemini 2.5 Pro Preview 0605

See all 919

• NanoGPT cut input 51.1% to $0.086 and output 9.3% to $1.40 on GLM 5.2 Thinking
• NanoGPT cut input 51.1% to $0.086 and output 9.3% to $1.40 on GLM 5.2
• NanoGPT cut input 32.4% to $0.050 and raised output 37.6% to $0.150 on Qwen3.5 9B
• NanoGPT cut input 78.7% to $0.010 and output 78.9% to $0.030 on Gemma 4 26B A4B Thinking
• NanoGPT cut input 78.7% to $0.010 and output 78.9% to $0.030 on Gemma 4 26B A4B
• NanoGPT cut input 75.6% to $0.022 and output 23.2% to $0.192 on DeepSeek V4.1 Flash Thinking
• NanoGPT cut input 75.6% to $0.022 and output 23.2% to $0.192 on DeepSeek V4.1 Flash
• SayGM raised input 26% to $1.30 and output 26% to $5.19 on o3
• SayGM cut input 2% to $1.70 and output 2% to $8.48 on kimi-k3
• SayGM cut input 10% to $0.649 and output 10% to $3.24 on GPT-6.1 Terra
• SayGM raised input 2.4% to $1.24 and output 2.4% to $6.18 on GPT-6.1 Sol
• SayGM raised input 2.4% to $1.24 and output 2.4% to $6.18 on GPT-6 Sol

See all 227

• LLM Gateway lowered Grok 4.1 Fast Reasoning's context window from 2M to 128K
• LLM Gateway lowered DeepSeek V4.1 Flash 1017's context window from 1.1M to 1.1M

• Cheaper Inference dropped qwen-image-2
• Cheaper Inference dropped longcat
• Cheaper Inference dropped laguna-s-2.1-free
• Cheaper Inference dropped kling-o3-4k
• Cheaper Inference dropped tts-qwen3-1-7b
• Cheaper Inference dropped tts-minimax-speech-02-hd
• Cheaper Inference dropped vidu-q3
• Cheaper Inference dropped nano-banana-2
• Cheaper Inference dropped venice-xai-stt-v1
• Cheaper Inference dropped nano-banana
• Cheaper Inference dropped venice-whisper-1
• Cheaper Inference dropped venice-wan-2.7-pro

See all 55

• qwen-image-2 is no longer served by any gateway we track
• longcat is no longer served by any gateway we track
• laguna-s-2.1-free is no longer served by any gateway we track
• kling-o3-4k is no longer served by any gateway we track
• tts-qwen3-1-7b is no longer served by any gateway we track
• tts-minimax-speech-02-hd is no longer served by any gateway we track
• vidu-q3 is no longer served by any gateway we track
• venice-xai-stt-v1 is no longer served by any gateway we track
• venice-whisper-1 is no longer served by any gateway we track
• venice-wan-2.7-pro is no longer served by any gateway we track
• venice-wai-illustrious is no longer served by any gateway we track
• venice-tts-1 is no longer served by any gateway we track

See all 48