← variables.md

2026-09-02

59 change(s) in the model market

added to the catalog

• Z.ai: GLM Flash Latest (~z-ai/glm-flash-latest) added to the catalog: 1,310,720 ctx, $0.075/M in / $0.25/M out
• Anthropic: Claude Fable 5.1 (anthropic/claude-fable-5.1) added to the catalog: 1,000,000 ctx, $10/M in / $50/M out
• Anthropic: Claude Fable 5.1 (batch) (anthropic/claude-fable-5.1:batch) added to the catalog: 1,000,000 ctx, $5/M in / $25/M out
• Inception: Mercury 2.5 Preview (inception/mercury-2.5-preview) added to the catalog: 260,000 ctx, $0.04/M in / $0.15/M out

removed from the catalog

• Anthropic: Claude Opus 4.7 (Fast) (anthropic/claude-opus-4.7-fast) removed from the catalog
• Anthropic: Claude Opus 4.8 (Fast) (anthropic/claude-opus-4.8-fast) removed from the catalog
• Claude Opus 5 (Fast) (anthropic/claude-opus-5-fast) removed from the catalog

model price changes

• deepseek/deepseek-chat-v3.1: prompt price $0.55/M → $0.25/M (−54.5%)
• deepseek/deepseek-chat-v3.1: completion price $1.65/M → $0.95/M (−42.4%)
• deepseek/deepseek-v4-flash: prompt price $0.081/M → $0.089/M (+9.5%)
• deepseek/deepseek-v4-flash: completion price $0.162/M → $0.177/M (+9.5%)
• deepseek/deepseek-v4-flash-vision-exp: prompt price $0.22/M → $0.44/M (+100.0%)
• deepseek/deepseek-v4-flash-vision-exp: completion price $0.66/M → $1.32/M (+100.0%)
• deepseek/deepseek-v4-pro: prompt price $1.6/M → $1.04/M (−34.9%)
• deepseek/deepseek-v4-pro: completion price $3.2/M → $2.08/M (−34.9%)
• deepseek/deepseek-v4-pro-0813: prompt price $0.66/M → $1.12/M (+69.0%)
• deepseek/deepseek-v4-pro-0813: completion price $1.98/M → $3.35/M (+69.0%)
• nvidia/nemotron-3-ultra-550b-a55b: prompt price $0.5/M → $0.625/M (+25.0%)
• nvidia/nemotron-3-ultra-550b-a55b: completion price $2.2/M → $3.13/M (+42.0%)
• qwen/qwen3.8-2.4t-a95b:batch: prompt price $2.5/M → $2/M (−20.0%)
• qwen/qwen3.8-2.4t-a95b:batch: completion price $6.25/M → $6/M (−4.0%)
• z-ai/glm-5.2: prompt price $1.19/M → $0.966/M (−18.8%)
• z-ai/glm-5.2: completion price $3.74/M → $3.04/M (−18.8%)

quantization changes

• GMICloud: deepseek/deepseek-v4-flash-0731 quantization fp8 → fp4

provider price changes

• Baidu: deepseek/deepseek-v4-flash prompt price $0.082/M → $0.14/M (+70.1%)
• Baidu: deepseek/deepseek-v4-flash completion price $0.165/M → $0.28/M (+70.1%)
• StreamLake: deepseek/deepseek-v4-flash prompt price $0.081/M → $0.089/M (+9.5%)
• StreamLake: deepseek/deepseek-v4-flash completion price $0.162/M → $0.177/M (+9.5%)
• DeepSeek: deepseek/deepseek-v4-flash-0731 prompt price $0.22/M → $0.44/M (+100.0%)
• DeepSeek: deepseek/deepseek-v4-flash-0731 completion price $0.66/M → $1.32/M (+100.0%)
• Reka: deepseek/deepseek-v4-flash-0731 prompt price $0.22/M → $0.11/M (−50.0%)
• StreamLake: deepseek/deepseek-v4-flash-0731 prompt price $0.22/M → $0.132/M (−40.0%)
• StreamLake: deepseek/deepseek-v4-flash-0731 completion price $0.66/M → $0.396/M (−40.0%)
• DeepSeek: deepseek/deepseek-v4-flash-vision-exp prompt price $0.22/M → $0.44/M (+100.0%)
• DeepSeek: deepseek/deepseek-v4-flash-vision-exp completion price $0.66/M → $1.32/M (+100.0%)
• Fireworks: deepseek/deepseek-v4-pro prompt price $1.74/M → $1.2/M (−31.0%)
• Fireworks: deepseek/deepseek-v4-pro completion price $3.48/M → $1.2/M (−65.5%)
• DeepSeek: deepseek/deepseek-v4-pro-0813 prompt price $0.66/M → $1.32/M (+100.0%)
• DeepSeek: deepseek/deepseek-v4-pro-0813 completion price $1.98/M → $3.96/M (+100.0%)
• StreamLake: moonshotai/kimi-k2.7-code prompt price $0.95/M → $0.712/M (−25.0%)
• StreamLake: moonshotai/kimi-k2.7-code completion price $4/M → $3/M (−25.0%)
• DeepInfra: stepfun/step-3.7-flash prompt price $0.2/M → $0.16/M (−20.0%)
• DeepInfra: stepfun/step-3.7-flash completion price $1.15/M → $0.92/M (−20.0%)
• Morph: z-ai/glm-5.3-flash prompt price $0.15/M → $0.13/M (−13.3%)
• Morph: z-ai/glm-5.3-flash completion price $0.42/M → $0.45/M (+7.1%)

provider context changes

• GMICloud: deepseek/deepseek-v4-flash-0731 context length 1,048,575 → 1,048,576

providers added

• Anthropic now serves anthropic/claude-opus-4.8 (unknown, 1,000,000 ctx, $10/M in / $50/M out)
• Anthropic now serves anthropic/claude-opus-5 (unknown, 1,000,000 ctx, $10/M in / $50/M out)
• Fireworks now serves deepseek/deepseek-v4-flash-vision-exp (unknown, 1,048,576 ctx, $0.22/M in / $0.66/M out)
• BaseTen now serves moonshotai/kimi-k2.7-code (fp4, 262,000 ctx, $0.95/M in / $4/M out)
• Wafer now serves moonshotai/kimi-k3 (unknown, 1,048,576 ctx, $3/M in / $15/M out)
• AkashML now serves openai/gpt-oss-20b (bf16, 131,072 ctx, $0.02/M in / $0.1/M out)
• Ionstream now serves qwen/qwen3.8-27b (fp8, 262,144 ctx, $0.35/M in / $2.55/M out)
• Inceptron now serves z-ai/glm-5.3 (fp4, 1,048,576 ctx, $1.4/M in / $4.4/M out)
• Makora now serves z-ai/glm-5.3 (fp4, 980,000 ctx, $1.35/M in / $4.4/M out)
• Sail Research now serves z-ai/glm-5.3 (fp8, 1,048,576 ctx, $1.4/M in / $4.4/M out)
• NextBit now serves z-ai/glm-5.3-flash (fp8, 1,048,576 ctx, $0.15/M in / $0.5/M out)

providers removed

• Parasail no longer serves meta/muse-glimmer-30b (was bf16)
• Parasail no longer serves xiaomi/mimo-v2.5 (was fp8)

openrouter catalog snapshots 2026-09-01 → 2026-09-02 · schema 1 · generated 2026-09-02T09:38:26.203Z