← variables.md

2026-08-27

93 change(s) in the model market

added to the catalog

• Ling 3.0 Flash Fin (free) (inclusionai/ling-3.0-flash-fin:free) added to the catalog: 262,144 ctx, $0/M in / $0/M out
• Qwen: Qwen3.8 Flash (qwen/qwen3.8-flash) added to the catalog: 1,000,000 ctx, $0.15/M in / $0.47/M out
• Z.ai: GLM 5.3 Flash (z-ai/glm-5.3-flash) added to the catalog: 1,310,720 ctx, $0.075/M in / $0.25/M out

removed from the catalog

• Ox Alpha (stealth/ox-alpha) removed from the catalog
• Z.ai: GLM 5.2 (batch) (z-ai/glm-5.2:batch) removed from the catalog

model price changes

• ~deepseek/deepseek-v4-flash-latest: completion price $0.075/M → $0.1/M (+33.3%)
• deepseek/deepseek-v4-flash: prompt price $0.089/M → $0.078/M (−12.0%)
• deepseek/deepseek-v4-flash: completion price $0.177/M → $0.156/M (−12.0%)
• moonshotai/kimi-k2.7-code: prompt price $0.67/M → $0.66/M (−1.5%)
• qwen/qwen3-235b-a22b-2507: prompt price $0.09/M → $0.087/M (−2.8%)
• qwen/qwen3-235b-a22b-2507: completion price $0.55/M → $0.35/M (−36.4%)
• qwen/qwen3.6-35b-a3b: prompt price $0.14/M → $0.1/M (−28.6%)
• qwen/qwen3.6-35b-a3b: completion price $1/M → $0.9/M (−10.0%)
• tencent/hy3: prompt price $0.132/M → $0.083/M (−37.5%)
• tencent/hy3: completion price $0.528/M → $0.33/M (−37.5%)
• z-ai/glm-4.6: prompt price $0.5/M → $0.43/M (−14.0%)
• z-ai/glm-4.6: completion price $2/M → $1.75/M (−12.5%)

quantization changes

• DeepInfra: xiaomi/mimo-v2.5 quantization bf16 → fp8

provider price changes

• Baidu: deepseek/deepseek-v4-flash prompt price $0.14/M → $0.079/M (−43.3%)
• Baidu: deepseek/deepseek-v4-flash completion price $0.28/M → $0.159/M (−43.3%)
• Mancer 2: deepseek/deepseek-v4-flash prompt price $0.14/M → $0.15/M (+7.1%)
• StreamLake: deepseek/deepseek-v4-flash prompt price $0.089/M → $0.078/M (−12.0%)
• StreamLake: deepseek/deepseek-v4-flash completion price $0.177/M → $0.156/M (−12.0%)
• Alibaba: deepseek/deepseek-v4-flash-0731 prompt price $0.352/M → $0.176/M (−50.0%)
• Alibaba: deepseek/deepseek-v4-flash-0731 completion price $1.06/M → $0.528/M (−50.0%)
• Baidu: deepseek/deepseek-v4-flash-0731 prompt price $0.14/M → $0.06/M (−57.2%)
• Baidu: deepseek/deepseek-v4-flash-0731 completion price $0.28/M → $0.12/M (−57.2%)
• Mancer 2: deepseek/deepseek-v4-flash-0731 prompt price $0.155/M → $0.15/M (−3.2%)
• Morph: deepseek/deepseek-v4-flash-0731 prompt price $0.079/M → $0.099/M (+25.0%)
• OpenInference: deepseek/deepseek-v4-flash-0731 completion price $0.075/M → $0.1/M (+33.3%)
• Phala: deepseek/deepseek-v4-flash-0731 prompt price $0.2/M → $0.22/M (+10.0%)
• Phala: deepseek/deepseek-v4-flash-0731 completion price $0.4/M → $0.66/M (+65.0%)
• Relace: deepseek/deepseek-v4-flash-0731 prompt price $0.06/M → $0.05/M (−16.7%)
• Relace: deepseek/deepseek-v4-flash-0731 completion price $0.12/M → $0.1/M (−16.7%)
• SiliconFlow: deepseek/deepseek-v4-flash-0731 prompt price $0.14/M → $0.22/M (+57.1%)
• SiliconFlow: deepseek/deepseek-v4-flash-0731 completion price $0.28/M → $0.66/M (+135.7%)
• StreamLake: deepseek/deepseek-v4-flash-0731 prompt price $0.247/M → $0.22/M (−10.9%)
• StreamLake: deepseek/deepseek-v4-flash-0731 completion price $0.741/M → $0.66/M (−10.9%)
• Baidu: deepseek/deepseek-v4-pro prompt price $1.69/M → $0.791/M (−53.2%)
• Baidu: deepseek/deepseek-v4-pro completion price $3.38/M → $1.58/M (−53.2%)
• Alibaba: deepseek/deepseek-v4-pro-0813 prompt price $1.12/M → $0.581/M (−48.2%)
• Alibaba: deepseek/deepseek-v4-pro-0813 completion price $3.37/M → $1.74/M (−48.2%)
• StreamLake: deepseek/deepseek-v4-pro-0813 prompt price $1.32/M → $0.66/M (−50.0%)
• StreamLake: deepseek/deepseek-v4-pro-0813 completion price $3.96/M → $1.98/M (−50.0%)
• Baidu: moonshotai/kimi-k2.6 prompt price $0.95/M → $0.548/M (−42.3%)
• Baidu: moonshotai/kimi-k2.6 completion price $4/M → $2.31/M (−42.3%)
• Decart: moonshotai/kimi-k2.6 prompt price $0.568/M → $0.548/M (−3.5%)
• Decart: moonshotai/kimi-k2.6 completion price $3.33/M → $2.31/M (−30.7%)
• Inceptron: moonshotai/kimi-k2.6 prompt price $0.57/M → $0.54/M (−5.3%)
• Inceptron: moonshotai/kimi-k2.7-code prompt price $0.67/M → $0.66/M (−1.5%)
• Mancer 2: openai/gpt-oss-120b prompt price $0.085/M → $0.08/M (−5.9%)
• AkashML: qwen/qwen3.6-35b-a3b prompt price $0.14/M → $0.1/M (−28.6%)
• AkashML: qwen/qwen3.6-35b-a3b completion price $1/M → $0.9/M (−10.0%)
• Baidu: tencent/hy3 prompt price $0.132/M → $0.132/M (−0.1%)
• Baidu: tencent/hy3 completion price $0.528/M → $0.527/M (−0.1%)
• Tencent: tencent/hy3 prompt price $0.132/M → $0.083/M (−37.5%)
• Tencent: tencent/hy3 completion price $0.528/M → $0.33/M (−37.5%)
• DeepInfra: xiaomi/mimo-v2.5 prompt price $0.4/M → $0.13/M (−67.5%)
• DeepInfra: xiaomi/mimo-v2.5 completion price $2/M → $0.65/M (−67.5%)
• Baidu: z-ai/glm-5.1 prompt price $1.4/M → $0.909/M (−35.1%)
• Baidu: z-ai/glm-5.1 completion price $4.4/M → $2.86/M (−35.1%)
• Baidu: z-ai/glm-5.2 prompt price $1.4/M → $0.419/M (−70.1%)
• Baidu: z-ai/glm-5.2 completion price $4.4/M → $1.32/M (−70.1%)
• StreamLake: z-ai/glm-5.2 prompt price $0.735/M → $0.42/M (−42.9%)
• StreamLake: z-ai/glm-5.2 completion price $2.31/M → $1.32/M (−42.9%)

provider context changes

• SambaNova: google/gemma-4-31b-it context length 131,072 → 262,144
• CoreWeave: meta-llama/llama-3.1-70b-instruct context length 128,000 → 131,072
• CoreWeave: meta-llama/llama-3.1-8b-instruct context length 128,000 → 131,072
• CoreWeave: z-ai/glm-5.2 context length 262,144 → 1,048,576

providers added

• NextBit now serves deepseek/deepseek-v4-flash-0731 (fp8, 1,048,576 ctx, $0.44/M in / $1.32/M out)
• NextBit now serves deepseek/deepseek-v4-pro (fp8, 1,048,576 ctx, $1.72/M in / $3.45/M out)
• Google now serves google/gemini-3.1-flash-lite (unknown, 1,048,576 ctx, $0.275/M in / $1.65/M out)
• Google now serves google/gemini-3.5-flash-lite (unknown, 1,048,576 ctx, $0.33/M in / $2.75/M out)
• Mistral now serves mistralai/mistral-large (unknown, 128,000 ctx, $2/M in / $6/M out)
• Mistral now serves mistralai/mistral-large (unknown, 128,000 ctx, $2.2/M in / $6.6/M out)
• Mistral now serves mistralai/mistral-large-2407 (unknown, 131,072 ctx, $2.2/M in / $6.6/M out)
• Mistral now serves mistralai/mistral-saba (unknown, 32,768 ctx, $0.22/M in / $0.66/M out)
• Mistral now serves mistralai/mixtral-8x22b-instruct (unknown, 65,536 ctx, $2.2/M in / $6.6/M out)
• Fireworks now serves moonshotai/kimi-k2.7-code (unknown, 262,144 ctx, $1.9/M in / $8/M out)
• Parasail now serves moonshotai/kimi-k3 (fp4, 1,048,576 ctx, $3/M in / $15/M out)
• Darkbloom now serves qwen/qwen3.5-35b-a3b (fp4, 262,144 ctx, $0.08/M in / $0.75/M out)
• Morph now serves qwen/qwen3.5-397b-a17b (fp4, 262,144 ctx, $0.5/M in / $3.5/M out)
• Morph now serves qwen/qwen3.6-27b (fp4, 131,072 ctx, $0.289/M in / $2.4/M out)
• Morph now serves qwen/qwen3.8-27b (fp4, 131,072 ctx, $0.289/M in / $2.4/M out)
• StreamLake now serves xiaomi/mimo-v2.5 (unknown, 1,000,000 ctx, $0.168/M in / $0.336/M out)
• AkashML now serves z-ai/glm-5.2 (fp8, 1,048,576 ctx, $0.77/M in / $2.42/M out)

providers removed

• Sail Research no longer serves deepseek/deepseek-v4-flash (was fp4)
• Io Net no longer serves deepseek/deepseek-v4-flash-0731 (was fp8)
• Sail Research no longer serves deepseek/deepseek-v4-pro-0813 (was fp4)
• Morph no longer serves minimax/minimax-m3 (was fp4)
• Sail Research no longer serves moonshotai/kimi-k3 (was fp4)
• Nebius no longer serves qwen/qwen3-next-80b-a3b-thinking (was fp8)
• Io Net no longer serves z-ai/glm-5.2 (was fp4)

openrouter catalog snapshots 2026-08-26 → 2026-08-27 · schema 1 · generated 2026-08-27T16:31:15.427Z