← variables.md

2026-08-20

64 change(s) in the model market

added to the catalog

• Z.ai: GLM Latest (~z-ai/glm-latest) added to the catalog: 1,048,576 ctx, $1.4/M in / $4.4/M out

removed from the catalog

• AI21: Jamba Large 1.7 (ai21/jamba-large-1.7) removed from the catalog
• Mancer: Weaver (alpha) (mancer/weaver) removed from the catalog

model price changes

• ~deepseek/deepseek-v4-flash-latest: prompt price $0.076/M → $0.065/M (−15.0%)
• ~deepseek/deepseek-v4-flash-latest: completion price $0.153/M → $0.14/M (−8.5%)
• deepseek/deepseek-chat-v3-0324: prompt price $0.27/M → $0.25/M (−7.4%)
• deepseek/deepseek-chat-v3-0324: completion price $1.12/M → $1/M (−10.7%)
• deepseek/deepseek-v4-flash: prompt price $0.083/M → $0.089/M (+7.3%)
• deepseek/deepseek-v4-flash: completion price $0.165/M → $0.177/M (+7.3%)
• deepseek/deepseek-v4-pro: prompt price $0.66/M → $1.44/M (+118.2%)
• deepseek/deepseek-v4-pro: completion price $1.98/M → $2.88/M (+45.5%)
• deepseek/deepseek-v4-pro-0813: prompt price $0.66/M → $1.19/M (+80.0%)
• deepseek/deepseek-v4-pro-0813: completion price $1.98/M → $3.56/M (+80.0%)
• minimax/minimax-m2.5: prompt price $0.22/M → $0.225/M (+2.3%)
• qwen/qwen3.5-35b-a3b: prompt price $0.225/M → $0.25/M (+11.1%)
• qwen/qwen3.5-35b-a3b: completion price $1.8/M → $1.25/M (−30.6%)
• qwen/qwen3.6-27b: prompt price $0.3/M → $0.6/M (+100.0%)
• qwen/qwen3.6-27b: completion price $2/M → $3.6/M (+80.0%)

model context changes

• qwen/qwen3.8-27b: context length 262,144 → 1,000,000

provider price changes

• GMICloud: deepseek/deepseek-v4-flash prompt price $0.084/M → $0.112/M (+33.3%)
• GMICloud: deepseek/deepseek-v4-flash completion price $0.168/M → $0.224/M (+33.3%)
• StreamLake: deepseek/deepseek-v4-flash prompt price $0.083/M → $0.089/M (+7.3%)
• StreamLake: deepseek/deepseek-v4-flash completion price $0.165/M → $0.177/M (+7.3%)
• Cloudflare: deepseek/deepseek-v4-flash-0731 prompt price $0.44/M → $0.08/M (−81.8%)
• Cloudflare: deepseek/deepseek-v4-flash-0731 completion price $1.32/M → $0.18/M (−86.4%)
• OpenInference: deepseek/deepseek-v4-flash-0731 prompt price $0.08/M → $0.065/M (−18.8%)
• OpenInference: deepseek/deepseek-v4-flash-0731 completion price $0.18/M → $0.14/M (−22.2%)
• Relace: deepseek/deepseek-v4-flash-0731 prompt price $0.105/M → $0.07/M (−33.3%)
• Relace: deepseek/deepseek-v4-flash-0731 completion price $0.21/M → $0.14/M (−33.3%)
• Sail Research: deepseek/deepseek-v4-flash-0731 prompt price $0.09/M → $0.065/M (−27.8%)
• GMICloud: deepseek/deepseek-v4-pro prompt price $0.696/M → $1.06/M (+51.7%)
• GMICloud: deepseek/deepseek-v4-pro completion price $1.39/M → $3.17/M (+127.6%)
• StreamLake: deepseek/deepseek-v4-pro prompt price $0.694/M → $1.04/M (+50.4%)
• StreamLake: deepseek/deepseek-v4-pro completion price $1.39/M → $2.09/M (+50.4%)
• AkashML: meta-llama/llama-3.3-70b-instruct prompt price $0.13/M → $0.2/M (+53.8%)
• AkashML: meta-llama/llama-3.3-70b-instruct completion price $0.4/M → $0.52/M (+30.0%)
• Decart: moonshotai/kimi-k2.6 prompt price $0.568/M → $0.587/M (+3.2%)
• Decart: moonshotai/kimi-k2.6 completion price $3.33/M → $2.47/M (−25.9%)
• Mancer 2: openai/gpt-oss-120b prompt price $0.085/M → $0.08/M (−5.9%)
• Chutes: qwen/qwen3.8-27b prompt price $0.45/M → $0.4/M (−11.1%)
• Chutes: qwen/qwen3.8-27b completion price $3.2/M → $3/M (−6.2%)
• Novita: z-ai/glm-5.2 prompt price $0.741/M → $0.665/M (−10.2%)
• Novita: z-ai/glm-5.2 completion price $2.33/M → $2.09/M (−10.2%)
• StreamLake: z-ai/glm-5.2 prompt price $0.734/M → $0.666/M (−9.2%)
• StreamLake: z-ai/glm-5.2 completion price $2.31/M → $2.09/M (−9.2%)

provider context changes

• Sail Research: deepseek/deepseek-v4-flash-0731 context length 262,144 → 1,048,576
• Together: moonshotai/kimi-k3 context length 1,000,000 → 1,048,576
• Nebius: qwen/qwen3-next-80b-a3b-thinking context length 8,000 → 128,000

providers added

• GMICloud now serves deepseek/deepseek-chat-v3-0324 (fp8, 163,840 ctx, $0.29/M in / $1.14/M out)
• Novita now serves inclusionai/ling-3.0-flash (unknown, 262,144 ctx, $0.042/M in / $0.126/M out)
• Mistral now serves mistralai/mistral-small-2603 (unknown, 262,144 ctx, $0.165/M in / $0.66/M out)
• GMICloud now serves moonshotai/kimi-k2.7-code (fp8, 262,144 ctx, $0.95/M in / $4/M out)
• Fireworks now serves moonshotai/kimi-k3 (unknown, 1,048,576 ctx, $3.3/M in / $16.5/M out)
• GMICloud now serves qwen/qwen3-235b-a22b-2507 (fp8, 262,144 ctx, $0.35/M in / $1.4/M out)
• Alibaba now serves qwen/qwen3.8-27b (unknown, 1,000,000 ctx, $0.55/M in / $3.3/M out)
• Parasail now serves qwen/qwen3.8-27b (fp8, 262,144 ctx, $0.45/M in / $3.2/M out)
• Reka now serves qwen/qwen3.8-27b (fp8, 262,144 ctx, $0.45/M in / $3.2/M out)
• Fireworks now serves z-ai/glm-5.2 (unknown, 1,048,576 ctx, $2.1/M in / $6.6/M out)

providers removed

• Inceptron no longer serves minimax/minimax-m2.5 (was fp8)
• Together no longer serves moonshotai/kimi-k2.6 (was unknown)
• AtlasCloud no longer serves moonshotai/kimi-k2.7-code (was int4)
• Nebius no longer serves nvidia/nemotron-3-super-120b-a12b (was fp4)
• AkashML no longer serves qwen/qwen3.5-35b-a3b (was fp8)
• AkashML no longer serves z-ai/glm-5.2 (was fp8)

openrouter catalog snapshots 2026-08-19 → 2026-08-20 · schema 1 · generated 2026-08-20T05:46:28.152Z