ModelMint

Price Changelog

LLM pricing changes tracked daily from LiteLLM.

RSSJSON

2026-09-23

bedrock/eu-west-3/mistral.mistral-large-2402-v1:0

input $/1M

$10.4→$5.2↓
bedrock/eu-west-3/mistral.mistral-large-2402-v1:0

output $/1M

$31.2→$15.6↓
bedrock/us-east-1/mistral.mistral-large-2402-v1:0

input $/1M

$8→$4↓
bedrock/us-east-1/mistral.mistral-large-2402-v1:0

output $/1M

$24→$12↓
bedrock/us-west-2/mistral.mistral-large-2402-v1:0

input $/1M

$8→$4↓
bedrock/us-west-2/mistral.mistral-large-2402-v1:0

output $/1M

$24→$12↓
cohere.command-text-v14

input $/1M

$1.5→$1↓
mistral.mistral-large-2402-v1:0

input $/1M

$8→$4↓
mistral.mistral-large-2402-v1:0

output $/1M

$24→$12↓
openrouter/deepseek/deepseek-v4-pro

input $/1M

$0.9553→$0.8923↓
openrouter/deepseek/deepseek-v4-pro

output $/1M

$1.9105→$1.7845↓
openrouter/deepseek/deepseek-v4-pro

cache read $/1M

$0.0796→$0.0744↓
openrouter/openai/gpt-oss-20b

input $/1M

$0.03→$0.018↓
openrouter/openai/gpt-oss-20b

output $/1M

$0.13→$0.09↓
openrouter/openai/gpt-oss-20b

max output tokens

117,964→32,768↓
eu.anthropic.claude-opus-4-5-20251101-v1:0

input $/1M

$5→$5.5↑
eu.anthropic.claude-opus-4-5-20251101-v1:0

output $/1M

$25→$27.5↑
eu.anthropic.claude-opus-4-5-20251101-v1:0

cache write $/1M

$6.25→$6.875↑
eu.anthropic.claude-opus-4-5-20251101-v1:0

cache read $/1M

$0.5→$0.55↑
gemini/gemini-robotics-er-2-streaming-preview

input $/1M

$2→$1↓
gemini/gemini-robotics-er-2-streaming-preview

output $/1M

$10→$5↓
openrouter/z-ai/glm-5.3

input $/1M

$0.84→$0.6538↓
openrouter/z-ai/glm-5.3

output $/1M

$2.64→$2.0548↓
openrouter/z-ai/glm-5.3

cache read $/1M

$0.156→$0.1214↓
openrouter/moonshotai/kimi-k2.7-code

output $/1M

$3.21→$3.3↑
openrouter/qwen/qwen3.6-27b

input $/1M

$0.3→$0.32↑
openrouter/qwen/qwen3.6-27b

output $/1M

$2→$2.7↑
openrouter/qwen/qwen3.6-27b

cache read $/1M

$0.03→$0.15↑
openrouter/qwen/qwen3.6-27b

max output tokens

65,536→262,140↑
openrouter/deepseek/deepseek-v4-flash

input $/1M

$0.0886→$0.049↓
openrouter/deepseek/deepseek-v4-flash

output $/1M

$0.1772→$0.098↓
openrouter/deepseek/deepseek-v4-flash

cache read $/1M

$0.0177→$0.0098↓
openrouter/qwen/qwen3-next-80b-a3b-thinking

max output tokens

32,768→235,929↑
openrouter/~anthropic/claude-opus-latest

input $/1M

$5→$4↓
openrouter/~anthropic/claude-opus-latest

output $/1M

$25→$20↓
openrouter/~anthropic/claude-opus-latest

cache write $/1M

$6.25→$5↓
openrouter/~anthropic/claude-opus-latest

cache read $/1M

$0.5→$0.2↓
openrouter/~deepseek/deepseek-flash-latest

input $/1M

$0.3→$0.12↓
openrouter/~deepseek/deepseek-flash-latest

output $/1M

$1.2→$0.48↓
openrouter/~deepseek/deepseek-flash-latest

cache read $/1M

$0.006→$0.0036↓
openrouter/~deepseek/deepseek-flash-latest

max output tokens

384,000→943,718↑
openrouter/~deepseek/deepseek-pro-latest

input $/1M

$1.32→$0.4↓
openrouter/~deepseek/deepseek-pro-latest

output $/1M

$3.96→$1.1999↓
openrouter/~deepseek/deepseek-pro-latest

cache read $/1M

$0.044→$0.0127↓
openrouter/~deepseek/deepseek-pro-latest

max output tokens

384,000→393,216↑
openrouter/~deepseek/deepseek-v4-flash-latest

input $/1M

$0.04→$0.03↓
openrouter/~deepseek/deepseek-v4-flash-latest

output $/1M

$0.64→$0.8↑
openrouter/~deepseek/deepseek-v4-flash-latest

cache read $/1M

$0.016→$0.008↓
openrouter/~moonshotai/kimi-latest

input $/1M

$3→$1.4989↓
openrouter/~moonshotai/kimi-latest

output $/1M

$15→$10.758↓
openrouter/~openai/gpt-luna-latest

input $/1M

$0.2→$0.1↓
openrouter/~openai/gpt-luna-latest

output $/1M

$1.2→$0.5↓
openrouter/~openai/gpt-luna-latest

cache write $/1M

$0.25→$0.125↓
openrouter/~openai/gpt-luna-latest

cache read $/1M

$0.02→$0.01↓
openrouter/~z-ai/glm-flash-latest

input $/1M

$0.15→$0.075↓
openrouter/~z-ai/glm-flash-latest

output $/1M

$0.5→$0.25↓
openrouter/~z-ai/glm-flash-latest

cache read $/1M

$0.05→$0.015↓
openrouter/~z-ai/glm-latest

input $/1M

$0.84→$0.6538↓
openrouter/~z-ai/glm-latest

output $/1M

$2.64→$2.0548↓
openrouter/~z-ai/glm-latest

cache read $/1M

$0.156→$0.1214↓
openrouter/aion-labs/aion-2.0

context window

131,072→1,048,576↑
openrouter/aion-labs/aion-3.0

context window

131,072→1,048,576↑
openrouter/aion-labs/aion-3.0-mini

context window

131,072→1,048,576↑
openrouter/moonshotai/kimi-k3:batch

input $/1M

$3→$2.28↓
openrouter/moonshotai/kimi-k3:batch

output $/1M

$15→$11.4↓
openrouter/moonshotai/kimi-k3:batch

cache read $/1M

$0.3→$0.228↓
openrouter/moonshotai/kimi-k3:batch

max output tokens

943,718→16,384↓
openrouter/z-ai/glm-5.3-flash:batch

input $/1M

$0.075→$0.06↓
openrouter/z-ai/glm-5.3-flash:batch

output $/1M

$0.25→$0.2↓
openrouter/z-ai/glm-5.3-flash:batch

max output tokens

943,718→131,072↓
openrouter/z-ai/glm-5.3:batch

input $/1M

$0.7→$0.72↑
openrouter/z-ai/glm-5.3:batch

output $/1M

$2.2→$2.4↑
openrouter/z-ai/glm-5.3:batch

cache read $/1M

$0.13→$0.12↓
openrouter/z-ai/glm-5.3:batch

max output tokens

943,718→131,072↓

2026-09-22

deepseek.v3.2

context window

163,840→164,000↑
deepseek.v3.2

max output tokens

163,840→8,000↓
google.gemma-3-12b-it

max output tokens

8,192→8,000↓
google.gemma-3-27b-it

max output tokens

8,192→8,000↓
google.gemma-3-4b-it

max output tokens

8,192→8,000↓
minimax.minimax-m2

context window

128,000→1,000,000↑
minimax.minimax-m2

max output tokens

8,192→8,000↓
minimax.minimax-m2.1

max output tokens

8,192→8,000↓
minimax.minimax-m2.5

context window

1,000,000→196,000↓
minimax.minimax-m2.5

max output tokens

8,192→8,000↓
mistral.devstral-2-123b

max output tokens

8,192→32,000↑
mistral.magistral-small-2509

max output tokens

8,192→40,000↑
mistral.ministral-3-14b-instruct

max output tokens

8,192→8,000↓
mistral.ministral-3-3b-instruct

max output tokens

8,192→8,000↓
mistral.ministral-3-8b-instruct

max output tokens

8,192→8,000↓
mistral.mistral-large-3-675b-instruct

context window

128,000→256,000↑
mistral.mistral-large-3-675b-instruct

max output tokens

8,192→32,000↑
moonshot.kimi-k2-thinking

context window

128,000→256,000↑
moonshot.kimi-k2-thinking

max output tokens

8,192→16,000↑
moonshotai.kimi-k2.5

context window

262,144→256,000↓
moonshotai.kimi-k2.5

max output tokens

262,144→16,000↓
nvidia.nemotron-nano-12b-v2

max output tokens

8,192→8,000↓
nvidia.nemotron-nano-9b-v2

max output tokens

8,192→8,000↓
nvidia.nemotron-nano-3-30b

context window

262,144→256,000↓
nvidia.nemotron-nano-3-30b

max output tokens

8,192→8,000↓
nvidia.nemotron-super-3-120b

max output tokens

32,768→32,000↓
openai.gpt-oss-safeguard-120b

max output tokens

8,192→16,000↑
openai.gpt-oss-safeguard-20b

max output tokens

8,192→16,000↑
openrouter/anthropic/claude-sonnet-4

context window

1,000,000→200,000↓
qwen.qwen3-next-80b-a3b

context window

128,000→256,000↑
qwen.qwen3-next-80b-a3b

max output tokens

8,192→8,000↓
qwen.qwen3-vl-235b-a22b

context window

128,000→256,000↑
qwen.qwen3-vl-235b-a22b

max output tokens

8,192→8,000↓
qwen.qwen3-coder-next

context window

262,144→256,000↓
qwen.qwen3-coder-next

max output tokens

8,192→16,000↑
together_ai/zai-org/GLM-4.6

context window

200,000→202,752↑
together_ai/zai-org/GLM-4.7

context window

200,000→202,752↑
together_ai/Qwen/Qwen3.7-Max

input $/1M

$2.5→$1.5↓
together_ai/Qwen/Qwen3.7-Max

output $/1M

$7.5→$4.5↓
together_ai/Qwen/Qwen3.7-Max

cache read $/1M

$0.5→$0.3↓
zai.glm-4.7

context window

200,000→203,000↑
zai.glm-4.7

max output tokens

128,000→4,000↓
zai.glm-4.7-flash

context window

200,000→203,000↑
zai.glm-4.7-flash

max output tokens

128,000→4,000↓
together_ai/Qwen/Qwen3.8-Flash

input $/1M

$0.15→$0.09↓
together_ai/Qwen/Qwen3.8-Flash

output $/1M

$0.47→$0.282↓
openrouter/z-ai/glm-5.3-flash

input $/1M

$0.09→$0.15↑
openrouter/z-ai/glm-5.3-flash

output $/1M

$0.3→$0.5↑
openrouter/z-ai/glm-5.3-flash

cache read $/1M

$0.018→$0.05↑
openrouter/z-ai/glm-5.3-flash

max output tokens

131,072→943,718↑
openrouter/z-ai/glm-5.3

input $/1M

$0.91→$0.84↓
openrouter/z-ai/glm-5.3

output $/1M

$2.86→$2.64↓
openrouter/z-ai/glm-5.3

cache read $/1M

$0.169→$0.156↓
openrouter/qwen/qwen3.8-27b

input $/1M

$0.2→$0.42↑
openrouter/qwen/qwen3.8-27b

output $/1M

$2.5→$3↑
openrouter/qwen/qwen3.8-27b

cache read $/1M

$0.05→$0.085↑
openrouter/deepseek/deepseek-v4-flash-0731

output $/1M

$0.16→$0.64↑
openrouter/moonshotai/kimi-k3

input $/1M

$1.7→$3↑
openrouter/moonshotai/kimi-k3

output $/1M

$8.5→$15↑
openrouter/moonshotai/kimi-k3

cache read $/1M

$0.17→$0.3↑
openrouter/nvidia/nemotron-3-nano-30b-a3b

input $/1M

$0.06→$0.05↓
openrouter/nvidia/nemotron-3-nano-30b-a3b

output $/1M

$0.24→$0.2↓
openrouter/meta-llama/llama-4-maverick

input $/1M

$0.2→$0.1875↓
openrouter/meta-llama/llama-4-maverick

output $/1M

$0.8→$0.6525↓
openrouter/meta-llama/llama-3.1-70b-instruct

input $/1M

$0.72→$0.4↓
openrouter/meta-llama/llama-3.1-70b-instruct

output $/1M

$0.72→$0.4↓
openrouter/meta-llama/llama-3.1-70b-instruct

max output tokens

8,192→16,384↑
openrouter/~deepseek/deepseek-v4-flash-latest

output $/1M

$0.16→$0.64↑
openrouter/~moonshotai/kimi-latest

input $/1M

$1.7→$3↑
openrouter/~moonshotai/kimi-latest

output $/1M

$8.5→$15↑
openrouter/~moonshotai/kimi-latest

cache read $/1M

$0.17→$0.3↑
openrouter/~x-ai/grok-latest

input $/1M

$2→$1.6↓
openrouter/~x-ai/grok-latest

output $/1M

$6→$4.8↓
openrouter/~x-ai/grok-latest

cache read $/1M

$0.5→$0.4↓
openrouter/~z-ai/glm-flash-latest

input $/1M

$0.09→$0.15↑
openrouter/~z-ai/glm-flash-latest

output $/1M

$0.3→$0.5↑
openrouter/~z-ai/glm-flash-latest

cache read $/1M

$0.018→$0.05↑
openrouter/~z-ai/glm-flash-latest

max output tokens

131,072→943,718↑
openrouter/~z-ai/glm-latest

input $/1M

$0.91→$0.84↓
openrouter/~z-ai/glm-latest

output $/1M

$2.86→$2.64↓
openrouter/~z-ai/glm-latest

cache read $/1M

$0.169→$0.156↓
openrouter/ibm-granite/granite-4.2-8b

input $/1M

$0.1→$0.06↓
openrouter/ibm-granite/granite-4.2-8b

output $/1M

$0.15→$0.25↑
openrouter/ibm-granite/granite-4.2-8b

cache read $/1M

$0.05→$0.015↓

2026-09-21

openrouter/deepseek/deepseek-v4-pro

input $/1M

$0.4223→$0.9553↑
openrouter/deepseek/deepseek-v4-pro

output $/1M

$0.8446→$1.9105↑
openrouter/deepseek/deepseek-v4-pro

cache read $/1M

$0.0352→$0.0796↑
openrouter/deepseek/deepseek-v4-flash-vision-exp

output $/1M

$0.6468→$0.66↑
openrouter/deepseek/deepseek-v4-flash-vision-exp

max output tokens

262,144→943,718↑
openrouter/z-ai/glm-5.3

input $/1M

$0.896→$0.91↑
openrouter/z-ai/glm-5.3

output $/1M

$2.816→$2.86↑
openrouter/qwen/qwen3.8-27b

input $/1M

$0.42→$0.2↓
openrouter/qwen/qwen3.8-27b

output $/1M

$3→$2.5↓
openrouter/qwen/qwen3.8-27b

cache read $/1M

$0.085→$0.05↓
openrouter/deepseek/deepseek-v4-flash-0731

output $/1M

$0.08→$0.16↑
openrouter/qwen/qwen3.6-35b-a3b

input $/1M

$0.1→$0.15↑
openrouter/qwen/qwen3.6-35b-a3b

output $/1M

$0.9→$1↑
openrouter/deepseek/deepseek-v4-flash

input $/1M

$0.0367→$0.0886↑
openrouter/deepseek/deepseek-v4-flash

output $/1M

$0.0734→$0.1772↑
openrouter/deepseek/deepseek-v4-flash

cache read $/1M

$0.0073→$0.0177↑
openrouter/meta-llama/llama-4-maverick

input $/1M

$0.1875→$0.2↑
openrouter/meta-llama/llama-4-maverick

output $/1M

$0.6525→$0.8↑
openrouter/meta-llama/llama-3.1-70b-instruct

input $/1M

$0.4→$0.72↑
openrouter/meta-llama/llama-3.1-70b-instruct

output $/1M

$0.4→$0.72↑
openrouter/meta-llama/llama-3.1-70b-instruct

max output tokens

16,384→8,192↓
openrouter/~deepseek/deepseek-flash-latest

input $/1M

$0.13→$0.3↑
openrouter/~deepseek/deepseek-flash-latest

output $/1M

$0.52→$1.2↑
openrouter/~deepseek/deepseek-flash-latest

cache read $/1M

$0.0026→$0.006↑
openrouter/~deepseek/deepseek-flash-latest

max output tokens

943,718→384,000↓
openrouter/~deepseek/deepseek-pro-latest

input $/1M

$0.5768→$1.32↑
openrouter/~deepseek/deepseek-pro-latest

output $/1M

$1.7305→$3.96↑
openrouter/~deepseek/deepseek-pro-latest

cache read $/1M

$0.0192→$0.044↑
openrouter/~deepseek/deepseek-v4-flash-latest

output $/1M

$0.08→$0.16↑
openrouter/~z-ai/glm-flash-latest

input $/1M

$0.075→$0.09↑
openrouter/~z-ai/glm-flash-latest

output $/1M

$0.25→$0.3↑
openrouter/~z-ai/glm-flash-latest

cache read $/1M

$0.015→$0.018↑
openrouter/~z-ai/glm-latest

input $/1M

$0.8442→$0.91↑
openrouter/~z-ai/glm-latest

output $/1M

$2.6532→$2.86↑
openrouter/~z-ai/glm-latest

cache read $/1M

$0.1568→$0.169↑
openrouter/ibm-granite/granite-4.2-8b

input $/1M

$0.06→$0.1↑
openrouter/ibm-granite/granite-4.2-8b

output $/1M

$0.25→$0.15↓
openrouter/ibm-granite/granite-4.2-8b

cache read $/1M

$0.015→$0.05↑
openrouter/meta/muse-glimmer-30b

input $/1M

$0.35→$0.3↓
openrouter/meta/muse-glimmer-30b

output $/1M

$1.5→$1.2↓
openrouter/meta/muse-glimmer-30b

max output tokens

117,964→16,384↓

2026-09-20

openrouter/deepseek/deepseek-v4-pro

input $/1M

$1.6→$0.4223↓
openrouter/deepseek/deepseek-v4-pro

output $/1M

$3.2→$0.8446↓
openrouter/deepseek/deepseek-v4-pro

cache read $/1M

$0.135→$0.0352↓
openrouter/deepseek/deepseek-v4-pro

max output tokens

393,216→384,000↓
openrouter/deepseek/deepseek-v4.1-flash

input $/1M

$0.15→$0.3↑
openrouter/deepseek/deepseek-v4.1-flash

output $/1M

$0.6→$1.2↑
openrouter/deepseek/deepseek-v4.1-flash

cache read $/1M

$0.003→$0.006↑
openrouter/deepseek/deepseek-v4-pro-0813

input $/1M

$0.5782→$1.32↑
openrouter/deepseek/deepseek-v4-pro-0813

output $/1M

$1.7345→$3.96↑
openrouter/deepseek/deepseek-v4-pro-0813

cache read $/1M

$0.0184→$0.044↑
openrouter/deepseek/deepseek-v4-pro-0813

max output tokens

393,216→384,000↓
openrouter/nvidia/nemotron-3.5-lightning

input $/1M

$0.08→$0.07↓
openrouter/nvidia/nemotron-3.5-lightning

max output tokens

131,072→235,929↑
openrouter/openai/gpt-oss-120b

max output tokens

117,964→65,536↓
openrouter/qwen/qwen3.5-35b-a3b

input $/1M

$0.1625→$0.3125↑
openrouter/qwen/qwen3.5-35b-a3b

output $/1M

$1.3→$1.25↓
openrouter/qwen/qwen3.5-35b-a3b

max output tokens

65,536→16,384↓
openrouter/deepseek/deepseek-v4-flash-vision-exp

max output tokens

943,718→262,144↓
openrouter/z-ai/glm-5.3

input $/1M

$0.91→$0.896↓
openrouter/z-ai/glm-5.3

output $/1M

$2.86→$2.816↓
openrouter/z-ai/glm-5.3

max output tokens

943,717→131,072↓
openrouter/qwen/qwen3.8-27b

input $/1M

$0.214→$0.42↑
openrouter/qwen/qwen3.8-27b

output $/1M

$2.55→$3↑
openrouter/qwen/qwen3.8-27b

cache read $/1M

$0.15→$0.085↓
openrouter/deepseek/deepseek-v4-flash-0731

input $/1M

$0.06→$0.04↓
openrouter/deepseek/deepseek-v4-flash-0731

output $/1M

$0.12→$0.08↓
openrouter/deepseek/deepseek-v4-flash-0731

cache read $/1M

$0.012→$0.016↑
openrouter/moonshotai/kimi-k3

input $/1M

$2.1→$1.7↓
openrouter/moonshotai/kimi-k3

output $/1M

$10.95→$8.5↓
openrouter/moonshotai/kimi-k3

cache read $/1M

$0.23→$0.17↓
openrouter/z-ai/glm-5.2

input $/1M

$0.5544→$0.6496↑
openrouter/z-ai/glm-5.2

output $/1M

$1.7424→$2.0416↑
openrouter/z-ai/glm-5.2

cache read $/1M

$0.103→$0.1206↑
openrouter/nvidia/nemotron-3-ultra-550b-a55b

input $/1M

$0.625→$0.6↓
openrouter/nvidia/nemotron-3-ultra-550b-a55b

output $/1M

$3.125→$2.4↓
openrouter/nvidia/nemotron-3-ultra-550b-a55b

cache read $/1M

$0.1875→$0.12↓
openrouter/nvidia/nemotron-3-ultra-550b-a55b

max output tokens

32,768→182,520↑
openrouter/deepseek/deepseek-v4-flash

input $/1M

$0.0498→$0.0367↓
openrouter/deepseek/deepseek-v4-flash

output $/1M

$0.0997→$0.0734↓
openrouter/qwen/qwen3.5-9b

max output tokens

235,929→32,768↓
openrouter/~deepseek/deepseek-flash-latest

input $/1M

$0.14→$0.13↓
openrouter/~deepseek/deepseek-flash-latest

output $/1M

$0.42→$0.52↑
openrouter/~deepseek/deepseek-flash-latest

max output tokens

393,216→943,718↑
openrouter/~deepseek/deepseek-pro-latest

max output tokens

393,216→384,000↓
openrouter/~deepseek/deepseek-v4-flash-latest

input $/1M

$0.055→$0.04↓
openrouter/~deepseek/deepseek-v4-flash-latest

output $/1M

$0.165→$0.08↓
openrouter/~deepseek/deepseek-v4-flash-latest

cache read $/1M

$0.0018→$0.016↑
openrouter/~deepseek/deepseek-v4-flash-latest

max output tokens

393,216→943,718↑
openrouter/~moonshotai/kimi-latest

input $/1M

$2.1→$1.7↓
openrouter/~moonshotai/kimi-latest

output $/1M

$10.95→$8.5↓
openrouter/~moonshotai/kimi-latest

cache read $/1M

$0.23→$0.17↓
openrouter/~z-ai/glm-latest

input $/1M

$0.9→$0.8442↓
openrouter/~z-ai/glm-latest

output $/1M

$2.805→$2.6532↓
openrouter/~z-ai/glm-latest

cache read $/1M

$0.1466→$0.1568↑
openrouter/~z-ai/glm-latest

max output tokens

235,929→131,072↓
openrouter/tencent/hy3

input $/1M

$0.0825→$0.132↑
openrouter/tencent/hy3

output $/1M

$0.33→$0.528↑
openrouter/tencent/hy3

cache read $/1M

$0.0206→$0.033↑

2026-09-19

openrouter/deepseek/deepseek-v4-pro

input $/1M

$0.9434→$1.6↑
openrouter/deepseek/deepseek-v4-pro

output $/1M

$1.8867→$3.2↑
openrouter/deepseek/deepseek-v4-pro

cache read $/1M

$0.0796→$0.135↑
openrouter/deepseek/deepseek-v4-pro-0813

input $/1M

$0.66→$0.5782↓
openrouter/deepseek/deepseek-v4-pro-0813

output $/1M

$1.98→$1.7345↓
replicate/google/gemini-2.5-flash

input $/1M

$2.5→$0.3↓
vercel_ai_gateway/google/gemini-2.5-pro

input $/1M

$2.5→$1.25↓
vertex_ai/google/gemma-4-26b-a4b-it-maas

context window

256,000→262,144↑
openrouter/z-ai/glm-5.3

input $/1M

$1.4→$0.91↓
openrouter/z-ai/glm-5.3

output $/1M

$4.4→$2.86↓
openrouter/z-ai/glm-5.3

cache read $/1M

$0.26→$0.169↓
openrouter/z-ai/glm-5.2

input $/1M

$0.4875→$0.5544↑
openrouter/z-ai/glm-5.2

output $/1M

$1.56→$1.7424↑
openrouter/z-ai/glm-5.2

cache read $/1M

$0.091→$0.103↑
openrouter/deepseek/deepseek-v4-flash

input $/1M

$0.0886→$0.0498↓
openrouter/deepseek/deepseek-v4-flash

output $/1M

$0.1772→$0.0997↓
openrouter/deepseek/deepseek-v4-flash

cache read $/1M

$0.0177→$0.01↓
openrouter/~deepseek/deepseek-flash-latest

input $/1M

$0.15→$0.14↓
openrouter/~deepseek/deepseek-flash-latest

output $/1M

$0.6→$0.42↓
openrouter/~deepseek/deepseek-flash-latest

cache read $/1M

$0.015→$0.0042↓
openrouter/~deepseek/deepseek-pro-latest

input $/1M

$0.66→$0.5782↓
openrouter/~deepseek/deepseek-pro-latest

output $/1M

$1.98→$1.7345↓
openrouter/~deepseek/deepseek-v4-flash-latest

output $/1M

$0.1767→$0.165↓
openrouter/~deepseek/deepseek-v4-flash-latest

cache read $/1M

$0.0088→$0.0018↓
openrouter/~z-ai/glm-latest

input $/1M

$0.8775→$0.9↑
openrouter/~z-ai/glm-latest

output $/1M

$2.97→$2.805↓
openrouter/~z-ai/glm-latest

cache read $/1M

$0.1755→$0.1466↓
openrouter/meta/muse-glimmer-30b

input $/1M

$0.3→$0.35↑
openrouter/meta/muse-glimmer-30b

output $/1M

$1.1→$1.5↑
openrouter/tencent/hy3

input $/1M

$0.132→$0.0825↓
openrouter/tencent/hy3

output $/1M

$0.528→$0.33↓
openrouter/tencent/hy3

cache read $/1M

$0.033→$0.0206↓

2026-09-18

azure_ai/FW-Kimi-K3

input $/1M

$3.3→$3↓
azure_ai/FW-Kimi-K3

output $/1M

$16.5→$15↓
azure_ai/FW-Kimi-K3

cache read $/1M

$0.33→$0.3↓
openrouter/anthropic/claude-sonnet-4.5

context window

200,000→1,000,000↑
openrouter/bytedance/ui-tars-1.5-7b

context window

131,072→128,000↓
openrouter/deepseek/deepseek-chat

input $/1M

$0.2574→$0.32↑
openrouter/deepseek/deepseek-chat

output $/1M

$1.0287→$0.89↓
openrouter/deepseek/deepseek-chat

context window

65,536→163,840↑
openrouter/deepseek/deepseek-chat

max output tokens

8,192→16,384↑
openrouter/deepseek/deepseek-chat-v3-0324

context window

65,536→163,840↑
openrouter/deepseek/deepseek-chat-v3-0324

max output tokens

8,192→147,456↑
openrouter/deepseek/deepseek-chat-v3.1

max output tokens

163,840→32,768↓
openrouter/deepseek/deepseek-v3.2-exp

max output tokens

163,840→65,536↓
openrouter/deepseek/deepseek-r1

context window

65,336→64,000↓
openrouter/deepseek/deepseek-r1

max output tokens

8,192→16,000↑
openrouter/deepseek/deepseek-r1-0528

context window

65,336→163,840↑
openrouter/deepseek/deepseek-r1-0528

max output tokens

8,192→32,768↑
openrouter/deepseek/deepseek-v4-pro

input $/1M

$1.6→$0.9434↓
openrouter/deepseek/deepseek-v4-pro

output $/1M

$3.2→$1.8867↓
openrouter/deepseek/deepseek-v4-pro

cache read $/1M

$0.135→$0.0796↓
openrouter/deepseek/deepseek-v4-pro

max output tokens

384,000→393,216↑
openrouter/deepseek/deepseek-v4-pro-0813

input $/1M

$0.5795→$0.66↑
openrouter/deepseek/deepseek-v4-pro-0813

output $/1M

$1.7384→$1.98↑
openrouter/deepseek/deepseek-v4-pro-0813

max output tokens

384,000→393,216↑
openrouter/google/gemini-2.5-pro

max output tokens

65,535→65,536↑
openrouter/google/gemini-3-flash-preview

max output tokens

65,535→65,536↑
openrouter/mancer/weaver

max output tokens

2,000→6,000↑
openrouter/minimax/minimax-m2

max output tokens

204,800→131,072↓
openrouter/mistralai/devstral-2512

max output tokens

65,536→209,715↑
openrouter/mistralai/ministral-3b-2512

max output tokens

131,072→104,857↓
openrouter/mistralai/ministral-8b-2512

max output tokens

262,144→209,715↓
openrouter/mistralai/ministral-14b-2512

max output tokens

262,144→209,715↓
openrouter/mistralai/mistral-large-2512

input $/1M

$0.5→$0.55↑
openrouter/mistralai/mistral-large-2512

output $/1M

$1.5→$1.65↑
openrouter/mistralai/mistral-large-2512

max output tokens

262,144→209,715↓
openrouter/mistralai/mistral-large

max output tokens

8,191→102,400↑
openrouter/mistralai/mistral-small-3.1-24b-instruct

context window

131,072→128,000↓
openrouter/mistralai/mistral-small-3.1-24b-instruct

max output tokens

131,072→102,400↓
openrouter/mistralai/mistral-small-3.2-24b-instruct

context window

128,000→256,000↑
openrouter/mistralai/mistral-small-3.2-24b-instruct

max output tokens

128,000→16,384↓
openrouter/mistralai/mixtral-8x22b-instruct

max output tokens

65,536→52,428↓
openrouter/moonshotai/kimi-k2.5

max output tokens

262,144→235,929↓
openrouter/openai/gpt-5.2-codex

context window

272,000→400,000↑
openrouter/openai/gpt-5

context window

272,000→400,000↑
openrouter/openai/gpt-5-mini

context window

272,000→400,000↑
openrouter/openai/gpt-5-nano

context window

272,000→400,000↑
openrouter/openai/gpt-5.2

context window

272,000→400,000↑
openrouter/openai/gpt-5.2-chat

max output tokens

16,384→32,000↑
openrouter/openai/gpt-5.2-pro

context window

272,000→400,000↑
openrouter/openai/gpt-oss-120b

input $/1M

$0.037→$0.15↑
openrouter/openai/gpt-oss-120b

output $/1M

$0.17→$0.6↑
openrouter/openai/gpt-oss-120b

max output tokens

32,768→117,964↑
openrouter/openai/gpt-oss-20b

max output tokens

32,768→117,964↑
openrouter/qwen/qwen-2.5-coder-32b-instruct

context window

33,792→32,768↓
openrouter/qwen/qwen-2.5-coder-32b-instruct

max output tokens

33,792→29,491↓
openrouter/qwen/qwen3-coder

context window

262,100→262,144↑
openrouter/qwen/qwen3-coder

max output tokens

262,100→65,536↓
openrouter/qwen/qwen3-coder-plus

context window

997,952→1,000,000↑
openrouter/qwen/qwen3-235b-a22b-2507

max output tokens

262,144→235,929↓
openrouter/qwen/qwen3-235b-a22b-thinking-2507

context window

262,144→131,072↓
openrouter/qwen/qwen3-235b-a22b-thinking-2507

max output tokens

262,144→117,964↓
openrouter/qwen/qwen3.5-397b-a17b

max output tokens

65,536→235,929↑
openrouter/undi95/remm-slerp-l2-13b

input $/1M

$0.45→$0.35↓
openrouter/undi95/remm-slerp-l2-13b

max output tokens

4,096→5,529↑
openrouter/z-ai/glm-4.6

context window

202,800→204,800↑
openrouter/z-ai/glm-4.6

max output tokens

131,000→16,384↓
openrouter/xiaomi/mimo-v2.5-pro

context window

1,048,576→1,050,000↑
openrouter/xiaomi/mimo-v2.5-pro

max output tokens

16,384→131,072↑
openrouter/xiaomi/mimo-v2.5

context window

1,048,576→1,050,000↑
openrouter/z-ai/glm-4.7

context window

202,752→204,800↑
openrouter/z-ai/glm-4.7

max output tokens

64,000→131,072↑
openrouter/z-ai/glm-4.7-flash

max output tokens

32,000→117,964↑
openrouter/z-ai/glm-5

context window

202,752→204,800↑
openrouter/z-ai/glm-5.1

context window

202,752→204,800↑
openrouter/z-ai/glm-5.1

max output tokens

65,535→128,000↑
openrouter/minimax/minimax-m2.1

context window

204,000→204,800↑
openrouter/minimax/minimax-m2.1

max output tokens

64,000→131,072↑
openrouter/minimax/minimax-m2.5

context window

196,608→204,800↑
openrouter/minimax/minimax-m2.5

max output tokens

65,536→128,000↑
wandb/openai/gpt-oss-120b

context window

131,072→131,000↓
wandb/openai/gpt-oss-20b

context window

131,072→131,000↓
wandb/meta-llama/Llama-3.1-8B-Instruct

context window

128,000→131,000↑
bedrock_mantle/xai.grok-4.3

context window

131,072→1,048,576↑
wandb/deepseek-ai/DeepSeek-V4-Flash

context window

1,048,576→1,049,000↑
wandb/deepseek-ai/DeepSeek-V4-Flash-0731

context window

262,144→262,000↓
wandb/deepseek-ai/DeepSeek-V4-Pro

context window

1,048,576→1,049,000↑
wandb/google/gemma-4-31B-it

context window

262,144→262,000↓
wandb/ibm-granite/granite-4.1-8b

context window

131,072→131,000↓
wandb/JetBrains/Mellum2-12B-A2.5B-Instruct

context window

131,072→131,000↓
wandb/meta-llama/Llama-3.1-70B-Instruct

context window

128,000→131,000↑
wandb/MiniMaxAI/MiniMax-M3

context window

262,144→262,000↓
wandb/moonshotai/Kimi-K2.7-Code

context window

262,144→262,000↓
wandb/moonshotai/Kimi-K2.6

context window

262,144→262,000↓
wandb/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B

input $/1M

$0.1→$0.07↓
wandb/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B

output $/1M

$0.25→$0.2↓
wandb/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B

cache read $/1M

$0.05→$0.04↓
wandb/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B

context window

262,144→262,000↓
wandb/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B

input $/1M

$0.75→$0.5↓
wandb/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B

output $/1M

$2.75→$2.15↓
wandb/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B

cache read $/1M

$0.15→$0.1↓
wandb/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B

context window

262,144→262,000↓
wandb/OpenPipe/Qwen3-14B-Instruct

context window

32,768→32,800↑
wandb/Qwen/Qwen3.8-27B

context window

262,144→262,000↓
wandb/Qwen/Qwen3.6-35B-A3B

context window

262,144→262,000↓
wandb/Qwen/Qwen3.6-27B

context window

262,144→262,000↓
wandb/Qwen/Qwen3.5-35B-A3B

context window

262,144→262,000↓
wandb/Qwen/Qwen3-30B-A3B-Instruct-2507

context window

262,144→262,000↓
wandb/zai-org/GLM-5.2

context window

262,144→1,049,000↑
bedrock_mantle/us-gov-west-1/xai.grok-4.3

context window

131,072→1,048,576↑
openrouter/google/gemini-3.5-flash

max output tokens

65,535→65,536↑
openrouter/openai/gpt-5.1

context window

272,000→400,000↑
openrouter/openai/gpt-5.3-codex

context window

272,000→400,000↑
openrouter/openai/gpt-5.4-mini

context window

272,000→400,000↑
openrouter/openai/gpt-5.4-nano

context window

272,000→400,000↑
openrouter/openai/gpt-5.6-luna

context window

922,000→1,050,000↑
openrouter/openai/gpt-5.6-terra

context window

922,000→1,050,000↑
openrouter/x-ai/grok-4.20

context window

1,000,000→2,000,000↑
openrouter/x-ai/grok-4.20

max output tokens

1,000,000→1,800,000↑
openrouter/x-ai/grok-4.20-multi-agent

context window

1,000,000→2,000,000↑
openrouter/x-ai/grok-4.20-multi-agent

max output tokens

1,000,000→1,800,000↑
openrouter/x-ai/grok-4.3

max output tokens

1,000,000→900,000↓
openrouter/x-ai/grok-4.5

max output tokens

500,000→450,000↓
openrouter/x-ai/grok-4.6

max output tokens

500,000→450,000↓
openrouter/x-ai/grok-build-0.1

max output tokens

256,000→230,400↓
openrouter/deepseek/deepseek-v4-flash-vision-exp

output $/1M

$0.66→$0.6468↓
openrouter/deepseek/deepseek-v4-flash-vision-exp

max output tokens

384,000→943,718↑
openrouter/z-ai/glm-5.3

cache read $/1M

$0.14→$0.26↑
openrouter/z-ai/glm-5.3

max output tokens

262,144→943,717↑
openrouter/qwen/qwen3.8-2.4t-a95b

max output tokens

262,144→131,072↓
openrouter/moonshotai/kimi-k3

input $/1M

$3→$2.1↓
openrouter/moonshotai/kimi-k3

output $/1M

$15→$10.95↓
openrouter/moonshotai/kimi-k3

cache read $/1M

$0.3→$0.23↓
openrouter/z-ai/glm-5.2

input $/1M

$1.4→$0.4875↓
openrouter/z-ai/glm-5.2

output $/1M

$4.4→$1.56↓
openrouter/z-ai/glm-5.2

cache read $/1M

$0.14→$0.091↓
openrouter/z-ai/glm-5.2:free

context window

256,000→32,768↓
openrouter/z-ai/glm-5.2:free

max output tokens

230,400→29,491↓
openrouter/deepseek/deepseek-v4-flash

output $/1M

$0.1708→$0.1772↑
openrouter/google/gemma-4-26b-a4b-it

max output tokens

16,384→235,929↑
openrouter/nvidia/nemotron-3-super-120b-a12b

input $/1M

$0.085→$0.08↓
openrouter/nvidia/nemotron-3-super-120b-a12b

output $/1M

$0.4→$0.45↑
openrouter/nvidia/nemotron-3-super-120b-a12b

context window

1,000,000→262,144↓
openrouter/nvidia/nemotron-3-super-120b-a12b

max output tokens

16,384→235,929↑
openrouter/nvidia/nemotron-3-nano-30b-a3b

input $/1M

$0.05→$0.06↑
openrouter/nvidia/nemotron-3-nano-30b-a3b

output $/1M

$0.2→$0.24↑
openrouter/moonshotai/kimi-k2-thinking

max output tokens

100,352→98,304↓
openrouter/qwen/qwen3-vl-30b-a3b-instruct

max output tokens

16,384→32,768↑
openrouter/qwen/qwen3-next-80b-a3b-instruct

max output tokens

235,929→16,384↓
openrouter/moonshotai/kimi-k2-0905

max output tokens

100,352→98,304↓
openrouter/moonshotai/kimi-k2

max output tokens

100,352→98,304↓
openrouter/meta-llama/llama-4-maverick

max output tokens

115,200→16,384↓

2026-09-17

azure/gpt-4o-2024-11-20

input $/1M

$2.75→$2.5↓
azure/gpt-4o-2024-11-20

output $/1M

$11→$10↓
azure/gpt-4o-mini-2024-07-18

input $/1M

$0.165→$0.15↓
azure/gpt-4o-mini-2024-07-18

output $/1M

$0.66→$0.6↓
azure/gpt-5.6-sol

input $/1M

$5→$4↓
azure/gpt-5.6-sol

output $/1M

$30→$20↓
azure/gpt-5.6-sol

cache write $/1M

$6.25→$5↓
azure/gpt-5.6-sol

cache read $/1M

$0.5→$0.4↓
azure/us/gpt-5.6-sol

input $/1M

$5.5→$4.4↓
azure/us/gpt-5.6-sol

output $/1M

$33→$22↓
azure/us/gpt-5.6-sol

cache write $/1M

$6.875→$5.5↓
azure/us/gpt-5.6-sol

cache read $/1M

$0.55→$0.44↓
azure/eu/gpt-5.6-sol

input $/1M

$5.5→$4.4↓
azure/eu/gpt-5.6-sol

output $/1M

$33→$22↓
azure/eu/gpt-5.6-sol

cache write $/1M

$6.875→$5.5↓
azure/eu/gpt-5.6-sol

cache read $/1M

$0.55→$0.44↓
azure/o1-mini

input $/1M

$1.21→$1.1↓
azure/o1-mini

output $/1M

$4.84→$4.4↓
azure/o1-mini

cache read $/1M

$0.605→$0.55↓
azure/us/o4-mini-2025-04-16

cache read $/1M

$0.31→$0.303↓
azure_ai/FW-GLM-5.2-Fast

input $/1M

$2.1→$2.31↑
azure_ai/FW-GLM-5.2-Fast

output $/1M

$6.6→$7.26↑
azure_ai/FW-GLM-5.2-Fast

cache read $/1M

$0.21→$0.231↑
azure_ai/FW-Inkling

input $/1M

$1→$1.1↑
azure_ai/FW-Inkling

output $/1M

$4.05→$4.46↑
azure_ai/FW-Inkling

cache read $/1M

$0.17→$0.19↑
azure_ai/FW-Nemotron-3-Ultra-NVFP4

input $/1M

$0.6→$0.66↑
azure_ai/FW-Nemotron-3-Ultra-NVFP4

output $/1M

$2.4→$2.64↑
azure_ai/FW-Nemotron-3-Ultra-NVFP4

cache read $/1M

$0.119→$0.13↑
azure_ai/Llama-4-Maverick-17B-128E-Instruct-FP8

input $/1M

$1.41→$0.25↓
azure_ai/Llama-4-Maverick-17B-128E-Instruct-FP8

output $/1M

$0.35→$1↑
azure_ai/Phi-4-mini-reasoning

output $/1M

$0.32→$0.3↓
claude-sonnet-4-5

context window

200,000→1,000,000↑
claude-sonnet-4-5-20250929

context window

200,000→1,000,000↑
gemini/gemini-robotics-er-2-preview

input $/1M

$2→$1↓
gemini/gemini-robotics-er-2-preview

output $/1M

$10→$5↓
gemini/gemini-robotics-er-2-preview

cache read $/1M

$0.2→$0.1↓
gemini/gemini-2.5-flash

max output tokens

65,535→65,536↑
gemini/gemini-2.5-flash-lite

max output tokens

65,535→65,536↑
gemini/gemini-flash-latest

input $/1M

$0.3→$0.75↑
gemini/gemini-flash-latest

output $/1M

$2.5→$3.75↑
gemini/gemini-flash-latest

cache read $/1M

$0.03→$0.075↑
gemini/gemini-flash-latest

max output tokens

65,535→65,536↑
gemini/gemini-flash-lite-latest

input $/1M

$0.1→$0.3↑
gemini/gemini-flash-lite-latest

output $/1M

$0.4→$2.5↑
gemini/gemini-flash-lite-latest

cache read $/1M

$0.01→$0.03↑
gemini/gemini-flash-lite-latest

max output tokens

65,535→65,536↑
gemini/gemini-2.5-pro

max output tokens

65,535→65,536↑
gemini/gemini-3-flash-preview

max output tokens

65,535→65,536↑
gemini/gemini-3.5-flash

max output tokens

65,535→65,536↑
gemini/gemini-2.5-pro-preview-tts

context window

1,048,576→8,192↓
gemini/gemini-2.5-pro-preview-tts

max output tokens

65,535→16,384↓
daybreak-blue-latest

context window

1,050,000→922,000↓
openrouter/deepseek/deepseek-chat

input $/1M

$0.32→$0.2574↓
openrouter/deepseek/deepseek-chat

output $/1M

$0.89→$1.0287↑
openrouter/deepseek/deepseek-v4-pro

input $/1M

$0.8599→$1.6↑
openrouter/deepseek/deepseek-v4-pro

output $/1M

$1.7198→$3.2↑
openrouter/deepseek/deepseek-v4-pro

cache read $/1M

$0.0717→$0.135↑
openrouter/gryphe/mythomax-l2-13b

input $/1M

$0.06→$0.08↑
openrouter/gryphe/mythomax-l2-13b

output $/1M

$0.06→$0.11↑
openrouter/mistralai/mistral-small-3.2-24b-instruct

input $/1M

$0.075→$0.0938↑
openrouter/mistralai/mistral-small-3.2-24b-instruct

output $/1M

$0.2→$0.25↑
openrouter/qwen/qwen3-235b-a22b-2507

input $/1M

$0.22→$0.0875↓
openrouter/qwen/qwen3-235b-a22b-2507

output $/1M

$0.88→$0.35↓
openrouter/qwen/qwen3.5-35b-a3b

input $/1M

$0.3125→$0.1625↓
openrouter/qwen/qwen3.5-35b-a3b

output $/1M

$1.25→$1.3↑
together_ai/Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8

context window

256,000→262,144↑
together_ai/deepseek-ai/DeepSeek-V3.1

context window

128,000→131,072↑
together_ai/zai-org/GLM-4.5-Air-FP8

context window

128,000→131,072↑
gemini-flash-latest

input $/1M

$0.3→$0.75↑
gemini-flash-latest

output $/1M

$2.5→$3.75↑
gemini-flash-latest

cache read $/1M

$0.03→$0.075↑
gemini-flash-latest

max output tokens

65,535→65,536↑
gemini-flash-lite-latest

input $/1M

$0.1→$0.3↑
gemini-flash-lite-latest

output $/1M

$0.4→$2.5↑
gemini-flash-lite-latest

cache read $/1M

$0.01→$0.03↑
gemini-flash-lite-latest

max output tokens

65,535→65,536↑
gemini-pro-latest

input $/1M

$1.25→$2↑
gemini-pro-latest

output $/1M

$10→$12↑
gemini-pro-latest

cache read $/1M

$0.125→$0.2↑
gemini-pro-latest

max output tokens

65,535→65,536↑
gemini/gemini-pro-latest

input $/1M

$1.25→$2↑
gemini/gemini-pro-latest

output $/1M

$10→$12↑
gemini/gemini-pro-latest

cache read $/1M

$0.125→$0.2↑
gemini/gemini-pro-latest

max output tokens

65,535→65,536↑
gemini/gemini-omni-1.1-flash

context window

131,072→1,048,576↑
openrouter/z-ai/glm-5.3-flash

input $/1M

$0.15→$0.09↓
openrouter/z-ai/glm-5.3-flash

output $/1M

$0.5→$0.3↓
openrouter/z-ai/glm-5.3-flash

cache read $/1M

$0.03→$0.018↓
openrouter/qwen/qwen3.8-27b

input $/1M

$0.42→$0.214↓
openrouter/qwen/qwen3.8-27b

output $/1M

$3→$2.55↓
openrouter/qwen/qwen3.8-27b

cache read $/1M

$0.085→$0.15↑
openrouter/deepseek/deepseek-v4-flash-0731

input $/1M

$0.065→$0.06↓
openrouter/deepseek/deepseek-v4-flash-0731

output $/1M

$0.18→$0.12↓
openrouter/deepseek/deepseek-v4-flash-0731

cache read $/1M

$0.016→$0.012↓
openrouter/moonshotai/kimi-k3

input $/1M

$2.1→$3↑
openrouter/moonshotai/kimi-k3

output $/1M

$10.53→$15↑
openrouter/moonshotai/kimi-k3

cache read $/1M

$0.235→$0.3↑
openrouter/z-ai/glm-5.2

input $/1M

$0.6→$1.4↑
openrouter/z-ai/glm-5.2

output $/1M

$2→$4.4↑
openrouter/z-ai/glm-5.2

cache read $/1M

$0.15→$0.14↓
openrouter/moonshotai/kimi-k2.7-code

output $/1M

$3.5→$3.21↓
openrouter/moonshotai/kimi-k2.7-code

cache read $/1M

$0.15→$0.18↑
openrouter/google/gemma-4-26b-a4b-it

input $/1M

$0.042→$0.09↑
openrouter/google/gemma-4-26b-a4b-it

output $/1M

$0.22→$0.3↑
openrouter/qwen/qwen3-vl-30b-a3b-instruct

input $/1M

$0.15→$0.13↓
openrouter/qwen/qwen3-vl-30b-a3b-instruct

output $/1M

$0.6→$0.52↓
openrouter/qwen/qwen3-30b-a3b-instruct-2507

input $/1M

$0.09→$0.0482↓
openrouter/qwen/qwen3-30b-a3b-instruct-2507

output $/1M

$0.3→$0.1931↓
openrouter/minimax/minimax-m1

input $/1M

$0.55→$0.4↓
openrouter/qwen/qwen3-14b

input $/1M

$0.2275→$0.12↓
openrouter/qwen/qwen3-14b

output $/1M

$0.91→$0.24↓
openrouter/meta-llama/llama-4-maverick

input $/1M

$0.2→$0.1875↓
openrouter/meta-llama/llama-4-maverick

output $/1M

$0.696→$0.6525↓

2026-09-15

fireworks_ai/accounts/fireworks/models/deepseek-v4-pro

input $/1M

$1.74→$1.2↓
fireworks_ai/accounts/fireworks/models/deepseek-v4-pro

output $/1M

$3.48→$1.2↓
fireworks_ai/accounts/fireworks/models/deepseek-v4-pro

cache read $/1M

$0.145→$0.6↑
fireworks_ai/deepseek-v4-pro

input $/1M

$1.74→$1.2↓
fireworks_ai/deepseek-v4-pro

output $/1M

$3.48→$1.2↓
fireworks_ai/deepseek-v4-pro

cache read $/1M

$0.145→$0.6↑
gemini-2.0-flash

input $/1M

$0.1→$0.15↑
gemini-2.0-flash

output $/1M

$0.4→$0.6↑

2026-09-13

openrouter/deepseek/deepseek-chat-v3.1

input $/1M

$0.2→$0.25↑
openrouter/deepseek/deepseek-chat-v3.1

output $/1M

$0.8→$0.95↑
openrouter/deepseek/deepseek-v3.2

max output tokens

163,840→65,536↓
openrouter/deepseek/deepseek-v4-pro

input $/1M

$1.32→$0.8599↓
openrouter/deepseek/deepseek-v4-pro

output $/1M

$3.96→$1.7198↓
openrouter/deepseek/deepseek-v4-pro-0813

input $/1M

$1.32→$0.5795↓
openrouter/deepseek/deepseek-v4-pro-0813

output $/1M

$3.96→$1.7384↓
openrouter/qwen/qwen3-235b-a22b-2507

input $/1M

$0.0875→$0.22↑
openrouter/qwen/qwen3-235b-a22b-2507

output $/1M

$0.35→$0.88↑
openrouter/qwen/qwen3.5-35b-a3b

input $/1M

$0.25→$0.3125↑
openrouter/qwen/qwen3.5-122b-a10b

input $/1M

$0.29→$0.26↓
openrouter/qwen/qwen3.5-122b-a10b

output $/1M

$2.4→$2.08↓
openrouter/z-ai/glm-4.6

input $/1M

$0.55→$0.43↓
openrouter/z-ai/glm-4.6

output $/1M

$2.2→$1.75↓
vertex_ai/mistral-small-2503

input $/1M

$1→$0.1↓
vertex_ai/mistral-small-2503

output $/1M

$3→$0.3↓
vertex_ai/mistral-small-2503@001

input $/1M

$1→$0.1↓
vertex_ai/mistral-small-2503@001

output $/1M

$3→$0.3↓
vertex_ai/openai/gpt-oss-20b-maas

input $/1M

$0.075→$0.07↓
vertex_ai/openai/gpt-oss-20b-maas

output $/1M

$0.3→$0.25↓
deepseek-v4-flash

input $/1M

$0.44→$0.3↓
deepseek-v4-flash

output $/1M

$1.32→$1.2↓
deepseek-v4-flash-vision-exp

input $/1M

$0.44→$0.3↓
deepseek-v4-flash-vision-exp

output $/1M

$1.32→$1.2↓
deepseek/deepseek-v4-flash

input $/1M

$0.44→$0.3↓
deepseek/deepseek-v4-flash

output $/1M

$1.32→$1.2↓
deepseek/deepseek-v4-flash-vision-exp

input $/1M

$0.44→$0.3↓
deepseek/deepseek-v4-flash-vision-exp

output $/1M

$1.32→$1.2↓
openrouter/z-ai/glm-5.3-flash

input $/1M

$0.075→$0.15↑
openrouter/z-ai/glm-5.3-flash

output $/1M

$0.25→$0.5↑
openrouter/z-ai/glm-5.3-flash

cache read $/1M

$0.015→$0.03↑
openrouter/moonshotai/kimi-k3

input $/1M

$3→$2.1↓
openrouter/moonshotai/kimi-k3

output $/1M

$15→$10.53↓
openrouter/moonshotai/kimi-k3

cache read $/1M

$0.3→$0.235↓
openrouter/z-ai/glm-5.2

input $/1M

$0.966→$0.6↓
openrouter/z-ai/glm-5.2

output $/1M

$3.036→$2↓
openrouter/z-ai/glm-5.2

cache read $/1M

$0.1932→$0.15↓
openrouter/moonshotai/kimi-k2.7-code

input $/1M

$0.66→$0.71↑
openrouter/moonshotai/kimi-k2.7-code

output $/1M

$3.4→$3.5↑
openrouter/moonshotai/kimi-k2.7-code

cache read $/1M

$0.18→$0.15↓
openrouter/deepseek/deepseek-v4-flash

output $/1M

$0.1756→$0.1708↓
openrouter/google/gemma-4-26b-a4b-it

input $/1M

$0.07→$0.042↓
openrouter/google/gemma-4-26b-a4b-it

output $/1M

$0.34→$0.22↓
openrouter/qwen/qwen3-next-80b-a3b-instruct

input $/1M

$0.1→$0.09↓
openrouter/qwen/qwen3-30b-a3b-instruct-2507

input $/1M

$0.0482→$0.09↑
openrouter/qwen/qwen3-30b-a3b-instruct-2507

output $/1M

$0.1931→$0.3↑
openrouter/qwen/qwen3-14b

input $/1M

$0.12→$0.2275↑
openrouter/qwen/qwen3-14b

output $/1M

$0.24→$0.91↑

2026-09-10

eu.anthropic.claude-3-5-haiku-20241022-v1:0

input $/1M

$0.25→$0.8↑
eu.anthropic.claude-3-5-haiku-20241022-v1:0

output $/1M

$1.25→$4↑
eu.anthropic.claude-3-5-haiku-20241022-v1:0

cache write $/1M

$0.3125→$1↑
eu.anthropic.claude-3-5-haiku-20241022-v1:0

cache read $/1M

$0.025→$0.08↑
vertex_ai/claude-haiku-4-5

max output tokens

8,192→64,000↑
vertex_ai/claude-haiku-4-5@20251001

max output tokens

8,192→64,000↑
vertex_ai/xai/grok-4.1-fast-non-reasoning

context window

2,000,000→128,000↓
vertex_ai/xai/grok-4.1-fast-non-reasoning

max output tokens

2,000,000→128,000↓
vertex_ai/xai/grok-4.1-fast-reasoning

context window

2,000,000→128,000↓
vertex_ai/xai/grok-4.1-fast-reasoning

max output tokens

2,000,000→128,000↓
vertex_ai/xai/grok-4.20-non-reasoning

input $/1M

$2→$1.25↓
vertex_ai/xai/grok-4.20-non-reasoning

output $/1M

$6→$2.5↓
vertex_ai/xai/grok-4.20-reasoning

input $/1M

$2→$1.25↓
vertex_ai/xai/grok-4.20-reasoning

output $/1M

$6→$2.5↓
bedrock_mantle/openai.gpt-oss-20b

input $/1M

$0.075→$0.07↓
bedrock_mantle/openai.gpt-oss-safeguard-20b

input $/1M

$0.075→$0.07↓
bedrock_mantle/openai.gpt-oss-safeguard-20b

output $/1M

$0.3→$0.2↓

2026-09-06

databricks/databricks-gpt-5-4

context window

272,000→922,000↑
nebius/google/gemma-3-27b-it

input $/1M

$0.06→$0.1↑
nebius/google/gemma-3-27b-it

output $/1M

$0.2→$0.3↑
nebius/google/gemma-3-27b-it

context window

128,000→110,000↓
nebius/google/gemma-3-27b-it

max output tokens

128,000→110,000↓
nebius/meta-llama/Llama-3.3-70B-Instruct

context window

128,000→131,072↑
nebius/meta-llama/Llama-3.3-70B-Instruct

max output tokens

128,000→131,072↑
nebius/Qwen/Qwen3-32B

context window

32,768→40,960↑
nebius/Qwen/Qwen3-32B

max output tokens

32,768→40,960↑
nebius/Qwen/Qwen2.5-VL-72B-Instruct

input $/1M

$0.13→$0.25↑
nebius/Qwen/Qwen2.5-VL-72B-Instruct

output $/1M

$0.4→$0.75↑
nebius/Qwen/Qwen2.5-VL-72B-Instruct

context window

131,072→32,000↓
nebius/Qwen/Qwen2.5-VL-72B-Instruct

max output tokens

131,072→32,000↓
openrouter/anthropic/claude-opus-4.5

max output tokens

32,000→64,000↑
openrouter/anthropic/claude-sonnet-4.5

context window

1,000,000→200,000↓
openrouter/anthropic/claude-sonnet-4.5

max output tokens

1,000,000→64,000↓
openrouter/anthropic/claude-haiku-4.5

max output tokens

200,000→64,000↓
openrouter/deepseek/deepseek-chat

input $/1M

$0.14→$0.32↑
openrouter/deepseek/deepseek-chat

output $/1M

$0.28→$0.89↑
openrouter/deepseek/deepseek-chat-v3-0324

input $/1M

$0.14→$0.25↑
openrouter/deepseek/deepseek-chat-v3-0324

output $/1M

$0.28→$1↑
openrouter/deepseek/deepseek-v3.2

input $/1M

$0.28→$0.269↓
openrouter/deepseek/deepseek-v3.2-exp

input $/1M

$0.2→$0.27↑
openrouter/deepseek/deepseek-v3.2-exp

output $/1M

$0.4→$0.41↑
openrouter/deepseek/deepseek-r1

input $/1M

$0.55→$0.7↑
openrouter/deepseek/deepseek-r1

output $/1M

$2.19→$2.5↑
openrouter/google/gemini-2.5-flash

max output tokens

8,192→65,535↑
openrouter/google/gemini-2.5-pro

max output tokens

8,192→65,535↑
openrouter/gryphe/mythomax-l2-13b

input $/1M

$1.875→$0.06↓
openrouter/gryphe/mythomax-l2-13b

output $/1M

$1.875→$0.06↓
openrouter/mancer/weaver

input $/1M

$5.625→$0.4↓
openrouter/mancer/weaver

output $/1M

$5.625→$0.75↓
openrouter/mistralai/devstral-2512

input $/1M

$0.15→$0.4↑
openrouter/mistralai/devstral-2512

output $/1M

$0.6→$2↑
openrouter/mistralai/mistral-large

input $/1M

$8→$2↓
openrouter/mistralai/mistral-large

output $/1M

$24→$6↓
openrouter/mistralai/mistral-small-3.1-24b-instruct

input $/1M

$0.1→$0.351↑
openrouter/mistralai/mistral-small-3.1-24b-instruct

output $/1M

$0.3→$0.555↑
openrouter/mistralai/mistral-small-3.2-24b-instruct

input $/1M

$0.1→$0.075↓
openrouter/mistralai/mistral-small-3.2-24b-instruct

output $/1M

$0.3→$0.2↓
openrouter/mistralai/mixtral-8x22b-instruct

input $/1M

$0.65→$2↑
openrouter/mistralai/mixtral-8x22b-instruct

output $/1M

$0.65→$6↑
openrouter/moonshotai/kimi-k2.5

input $/1M

$0.6→$0.45↓
openrouter/moonshotai/kimi-k2.5

output $/1M

$3→$2.25↓
openrouter/moonshotai/kimi-k2.5

cache read $/1M

$0.1→$0.07↓
openrouter/nvidia/nemotron-3.5-lightning

input $/1M

$0.05→$0.08↑
openrouter/openai/gpt-3.5-turbo

input $/1M

$1.5→$0.5↓
openrouter/openai/gpt-3.5-turbo

output $/1M

$2→$1.5↓
openrouter/openai/gpt-4o

max output tokens

4,096→16,384↑
openrouter/openai/gpt-oss-120b

input $/1M

$0.18→$0.037↓
openrouter/openai/gpt-oss-120b

output $/1M

$0.8→$0.17↓
openrouter/openai/gpt-oss-20b

input $/1M

$0.02→$0.03↑
openrouter/openai/gpt-oss-20b

output $/1M

$0.1→$0.13↑
openrouter/openai/o3-mini

context window

128,000→200,000↑
openrouter/openai/o3-mini

max output tokens

65,536→100,000↑
openrouter/openai/o3-mini-high

context window

128,000→200,000↑
openrouter/openai/o3-mini-high

max output tokens

65,536→100,000↑
openrouter/qwen/qwen-2.5-coder-32b-instruct

input $/1M

$0.18→$0.66↑
openrouter/qwen/qwen-2.5-coder-32b-instruct

output $/1M

$0.18→$1↑
openrouter/qwen/qwen3-coder

input $/1M

$0.22→$0.3↑
openrouter/qwen/qwen3-coder

output $/1M

$0.95→$1↑
openrouter/qwen/qwen3-coder-plus

input $/1M

$1→$0.65↓
openrouter/qwen/qwen3-coder-plus

output $/1M

$5→$3.25↓
openrouter/qwen/qwen3-235b-a22b-2507

input $/1M

$0.071→$0.0875↑
openrouter/qwen/qwen3-235b-a22b-2507

output $/1M

$0.1→$0.35↑
openrouter/qwen/qwen3-235b-a22b-thinking-2507

input $/1M

$0.11→$0.23↑
openrouter/qwen/qwen3-235b-a22b-thinking-2507

output $/1M

$0.6→$2.3↑
openrouter/qwen/qwen3.5-35b-a3b

output $/1M

$2→$1.25↓
openrouter/qwen/qwen3.5-27b

input $/1M

$0.3→$0.195↓
openrouter/qwen/qwen3.5-27b

output $/1M

$2.4→$1.56↓
openrouter/qwen/qwen3.5-122b-a10b

input $/1M

$0.4→$0.29↓
openrouter/qwen/qwen3.5-122b-a10b

output $/1M

$2→$2.4↑
openrouter/qwen/qwen3.5-flash-02-23

input $/1M

$0.1→$0.065↓
openrouter/qwen/qwen3.5-flash-02-23

output $/1M

$0.4→$0.26↓
openrouter/qwen/qwen3.5-plus-02-15

input $/1M

$0.4→$0.26↓
openrouter/qwen/qwen3.5-plus-02-15

output $/1M

$2.4→$1.56↓
openrouter/qwen/qwen3.5-397b-a17b

input $/1M

$0.6→$0.55↓
openrouter/qwen/qwen3.5-397b-a17b

output $/1M

$3.6→$3.5↓
openrouter/undi95/remm-slerp-l2-13b

input $/1M

$1.875→$0.45↓
openrouter/undi95/remm-slerp-l2-13b

output $/1M

$1.875→$0.65↓
openrouter/z-ai/glm-4.6

input $/1M

$0.4→$0.55↑
openrouter/z-ai/glm-4.6

output $/1M

$1.75→$2.2↑
openrouter/xiaomi/mimo-v2.5-pro

input $/1M

$1→$0.435↓
openrouter/xiaomi/mimo-v2.5-pro

output $/1M

$3→$0.87↓
openrouter/xiaomi/mimo-v2.5-pro

cache read $/1M

$0.2→$0.0036↓
openrouter/xiaomi/mimo-v2.5

input $/1M

$0.4→$0.14↓
openrouter/xiaomi/mimo-v2.5

output $/1M

$2→$0.28↓
openrouter/xiaomi/mimo-v2.5

cache read $/1M

$0.08→$0.0028↓
openrouter/z-ai/glm-4.7

output $/1M

$1.5→$1.75↑
openrouter/z-ai/glm-4.7-flash

input $/1M

$0.07→$0.06↓
openrouter/z-ai/glm-5

input $/1M

$0.8→$0.6↓
openrouter/z-ai/glm-5

output $/1M

$2.56→$1.92↓
openrouter/z-ai/glm-5.1

input $/1M

$1.05→$0.966↓
openrouter/z-ai/glm-5.1

output $/1M

$3.5→$3.036↓
openrouter/z-ai/glm-5.1

cache read $/1M

$0.525→$0.1794↓
openrouter/minimax/minimax-m2.1

input $/1M

$0.27→$0.3↑
openrouter/minimax/minimax-m2.5

input $/1M

$0.3→$0.27↓
openrouter/minimax/minimax-m2.5

output $/1M

$1.1→$1.08↓
openrouter/minimax/minimax-m2.5

cache read $/1M

$0.15→$0.027↓
qwen.qwen3-coder-480b-a35b-v1:0

input $/1M

$0.22→$0.45↑
wandb/Qwen/Qwen3-235B-A22B-Instruct-2507

input $/1M

$10000→$0.1↓
wandb/Qwen/Qwen3-235B-A22B-Instruct-2507

output $/1M

$10000→$0.1↓
wandb/Qwen/Qwen3-235B-A22B-Thinking-2507

input $/1M

$10000→$0.1↓
wandb/Qwen/Qwen3-235B-A22B-Thinking-2507

output $/1M

$10000→$0.1↓
wandb/deepseek-ai/DeepSeek-R1-0528

input $/1M

$135000→$1.35↓
wandb/deepseek-ai/DeepSeek-R1-0528

output $/1M

$540000→$5.4↓
wandb/deepseek-ai/DeepSeek-V3-0324

input $/1M

$114000→$1.14↓
wandb/deepseek-ai/DeepSeek-V3-0324

output $/1M

$275000→$2.75↓
wandb/meta-llama/Llama-4-Scout-17B-16E-Instruct

input $/1M

$17000→$0.17↓
wandb/meta-llama/Llama-4-Scout-17B-16E-Instruct

output $/1M

$66000→$0.66↓
watsonx/bigscience/mt0-xxl-13b

input $/1M

$500→$1.908↓
watsonx/bigscience/mt0-xxl-13b

output $/1M

$2000→$1.908↓
watsonx/bigscience/mt0-xxl-13b

context window

8,192→4,096↓
watsonx/bigscience/mt0-xxl-13b

max output tokens

8,192→4,096↓
watsonx/ibm/granite-4-h-small

output $/1M

$0.25→$0.265↑
watsonx/ibm/granite-4-h-small

context window

20,480→131,072↑
watsonx/ibm/granite-4-h-small

max output tokens

20,480→131,072↑
watsonx/meta-llama/llama-3-3-70b-instruct

input $/1M

$0.71→$0.7526↑
watsonx/meta-llama/llama-3-3-70b-instruct

output $/1M

$0.71→$0.7526↑
watsonx/meta-llama/llama-3-3-70b-instruct

context window

128,000→131,072↑
watsonx/meta-llama/llama-3-3-70b-instruct

max output tokens

128,000→131,072↑
watsonx/meta-llama/llama-4-maverick-17b

input $/1M

$0.35→$0.371↑
watsonx/meta-llama/llama-4-maverick-17b

output $/1M

$1.4→$1.484↑
watsonx/meta-llama/llama-4-maverick-17b

context window

128,000→131,072↑
watsonx/meta-llama/llama-4-maverick-17b

max output tokens

128,000→8,192↓
watsonx/mistralai/mistral-small-3-1-24b-instruct-2503

input $/1M

$0.1→$0.106↑
watsonx/mistralai/mistral-small-3-1-24b-instruct-2503

output $/1M

$0.3→$0.318↑
watsonx/mistralai/mistral-small-3-1-24b-instruct-2503

context window

32,000→131,072↑
watsonx/mistralai/mistral-small-3-1-24b-instruct-2503

max output tokens

32,000→16,384↓
watsonx/openai/gpt-oss-120b

input $/1M

$0.15→$0.159↑
watsonx/openai/gpt-oss-120b

output $/1M

$0.6→$0.636↑
watsonx/openai/gpt-oss-120b

context window

8,192→131,072↑
watsonx/openai/gpt-oss-120b

max output tokens

8,192→131,072↑

2026-09-04

bedrock/us-gov-west-1/meta.llama3-8b-instruct-v1:0

output $/1M

$2.65→$0.6↓
mistral/magistral-medium-latest

input $/1M

$2→$1.5↓
mistral/magistral-medium-latest

output $/1M

$5→$7.5↑
mistral/magistral-medium-latest

context window

40,000→262,144↑
mistral/magistral-medium-latest

max output tokens

40,000→262,144↑
mistral/magistral-small-latest

input $/1M

$0.5→$0.15↓
mistral/magistral-small-latest

output $/1M

$1.5→$0.6↓
mistral/magistral-small-latest

context window

40,000→262,144↑
mistral/magistral-small-latest

max output tokens

40,000→262,144↑
mistral/mistral-medium

input $/1M

$2.7→$1.5↓
mistral/mistral-medium

output $/1M

$8.1→$7.5↓
mistral/mistral-medium

context window

32,000→262,144↑
mistral/mistral-medium

max output tokens

8,191→262,144↑
together_ai/Qwen/Qwen3.8-2.4T-A95B

input $/1M

$2.5→$2↓
together_ai/Qwen/Qwen3.8-2.4T-A95B

output $/1M

$6.25→$6↓
together_ai/Qwen/Qwen3.8-2.4T-A95B

cache read $/1M

$0.5→$0.25↓

2026-09-02

fireworks_ai/deepseek-v4-flash-0731

input $/1M

$0.14→$0.22↑
fireworks_ai/deepseek-v4-flash-0731

output $/1M

$0.28→$0.66↑
fireworks_ai/deepseek-v4-flash-0731

cache read $/1M

$0.028→$0.007↓

2026-08-30

gemini/gemini-omni-flash-preview

context window

1,048,576→131,072↓
gemini/gemini-omni-flash-preview

max output tokens

65,535→65,536↑
together_ai/meta-llama/Llama-3.3-70B-Instruct-Turbo

max output tokens

131,072→—
together_ai/openai/gpt-oss-120b

max output tokens

131,072→—
together_ai/openai/gpt-oss-20b

context window

128,000→131,072↑
together_ai/zai-org/GLM-4.6

max output tokens

200,000→—
together_ai/zai-org/GLM-4.7

max output tokens

200,000→—
together_ai/moonshotai/Kimi-K2.5

max output tokens

256,000→—
together_ai/MiniMaxAI/MiniMax-M3

max output tokens

524,288→—
together_ai/Prism-ML/Ternary-Bonsai-27B

max output tokens

262,144→—
together_ai/Qwen/Qwen3.5-9B

max output tokens

262,144→—
together_ai/Qwen/Qwen3.6-Plus

max output tokens

1,000,000→—
together_ai/Qwen/Qwen3.7-Max

max output tokens

1,000,000→—
together_ai/Qwen/Qwen3.7-Plus

max output tokens

1,000,000→—
together_ai/Qwen/Qwen3.8-2.4T-A95B

max output tokens

1,010,000→—
together_ai/arize-ai/qwen-2-1.5b-instruct

max output tokens

32,768→—
together_ai/deepseek-ai/DeepSeek-V4-Flash-0731

max output tokens

1,048,576→—
together_ai/deepseek-ai/DeepSeek-V4-Pro

max output tokens

512,000→—
together_ai/deepseek-ai/DeepSeek-V4-Pro-0813

max output tokens

1,048,576→—
together_ai/google/gemma-3n-E4B-it

max output tokens

32,768→—
together_ai/google/gemma-4-31B-it

max output tokens

262,144→—
together_ai/meta-llama/Llama-Guard-4-12B

max output tokens

1,048,576→—
together_ai/meta-models/Muse-Glimmer-30B

max output tokens

131,072→—
together_ai/moonshotai/Kimi-K2.7-Code

max output tokens

262,144→—
together_ai/moonshotai/Kimi-K3

max output tokens

1,048,576→—
together_ai/nvidia/nemotron-3-ultra-550b-a55b

max output tokens

512,288→—
together_ai/pearl-ai/gemma-4-31b-it

max output tokens

262,144→—
together_ai/thinkingmachines/Inkling

max output tokens

524,288→—
together_ai/thinkingmachines/Inkling-Small

max output tokens

524,288→—
together_ai/zai-org/GLM-5.2

max output tokens

1,048,575→128,000↓
together_ai/zai-org/GLM-5.3-Flash

max output tokens

1,048,575→128,000↓
xai/grok-3

input $/1M

$3→$1.25↓
xai/grok-3

output $/1M

$15→$2.5↓
xai/grok-3

cache read $/1M

$0.75→$0.2↓
xai/grok-3-beta

input $/1M

$3→$1.25↓
xai/grok-3-beta

output $/1M

$15→$2.5↓
xai/grok-3-beta

cache read $/1M

$0.75→$0.2↓
xai/grok-3-fast-beta

input $/1M

$5→$1.25↓
xai/grok-3-fast-beta

output $/1M

$25→$2.5↓
xai/grok-3-fast-beta

cache read $/1M

$1.25→$0.2↓
xai/grok-3-fast-latest

input $/1M

$5→$1.25↓
xai/grok-3-fast-latest

output $/1M

$25→$2.5↓
xai/grok-3-fast-latest

cache read $/1M

$1.25→$0.2↓
xai/grok-3-latest

input $/1M

$3→$1.25↓
xai/grok-3-latest

output $/1M

$15→$2.5↓
xai/grok-3-latest

cache read $/1M

$0.75→$0.2↓
xai/grok-3-mini

input $/1M

$0.3→$1.25↑
xai/grok-3-mini

output $/1M

$0.5→$2.5↑
xai/grok-3-mini

cache read $/1M

$0.075→$0.2↑
xai/grok-3-mini-beta

input $/1M

$0.3→$1.25↑
xai/grok-3-mini-beta

output $/1M

$0.5→$2.5↑
xai/grok-3-mini-beta

cache read $/1M

$0.075→$0.2↑
xai/grok-3-mini-fast

input $/1M

$0.6→$1.25↑
xai/grok-3-mini-fast

output $/1M

$4→$2.5↓
xai/grok-3-mini-fast

cache read $/1M

$0.15→$0.2↑
xai/grok-3-mini-fast-beta

input $/1M

$0.6→$1.25↑
xai/grok-3-mini-fast-beta

output $/1M

$4→$2.5↓
xai/grok-3-mini-fast-beta

cache read $/1M

$0.15→$0.2↑
xai/grok-3-mini-fast-latest

input $/1M

$0.6→$1.25↑
xai/grok-3-mini-fast-latest

output $/1M

$4→$2.5↓
xai/grok-3-mini-fast-latest

cache read $/1M

$0.15→$0.2↑
xai/grok-3-mini-latest

input $/1M

$0.3→$1.25↑
xai/grok-3-mini-latest

output $/1M

$0.5→$2.5↑
xai/grok-3-mini-latest

cache read $/1M

$0.075→$0.2↑
xai/grok-4

input $/1M

$3→$1.25↓
xai/grok-4

output $/1M

$15→$2.5↓
xai/grok-4-fast-reasoning

input $/1M

$0.2→$1.25↑
xai/grok-4-fast-reasoning

output $/1M

$0.5→$2.5↑
xai/grok-4-fast-reasoning

cache read $/1M

$0.05→$0.2↑
xai/grok-4-fast-non-reasoning

input $/1M

$0.2→$1.25↑
xai/grok-4-fast-non-reasoning

output $/1M

$0.5→$2.5↑
xai/grok-4-fast-non-reasoning

cache read $/1M

$0.05→$0.2↑
xai/grok-4-0709

input $/1M

$3→$1.25↓
xai/grok-4-0709

output $/1M

$15→$2.5↓
xai/grok-4-latest

input $/1M

$3→$1.25↓
xai/grok-4-latest

output $/1M

$15→$2.5↓
xai/grok-4-1-fast

input $/1M

$0.2→$1.25↑
xai/grok-4-1-fast

output $/1M

$0.5→$2.5↑
xai/grok-4-1-fast

cache read $/1M

$0.05→$0.2↑
xai/grok-4-1-fast-reasoning

input $/1M

$0.2→$1.25↑
xai/grok-4-1-fast-reasoning

output $/1M

$0.5→$2.5↑
xai/grok-4-1-fast-reasoning

cache read $/1M

$0.05→$0.2↑
xai/grok-4-1-fast-reasoning-latest

input $/1M

$0.2→$1.25↑
xai/grok-4-1-fast-reasoning-latest

output $/1M

$0.5→$2.5↑
xai/grok-4-1-fast-reasoning-latest

cache read $/1M

$0.05→$0.2↑
xai/grok-4-1-fast-non-reasoning

input $/1M

$0.2→$1.25↑
xai/grok-4-1-fast-non-reasoning

output $/1M

$0.5→$2.5↑
xai/grok-4-1-fast-non-reasoning

cache read $/1M

$0.05→$0.2↑
xai/grok-4-1-fast-non-reasoning-latest

input $/1M

$0.2→$1.25↑
xai/grok-4-1-fast-non-reasoning-latest

output $/1M

$0.5→$2.5↑
xai/grok-4-1-fast-non-reasoning-latest

cache read $/1M

$0.05→$0.2↑

2026-08-28

deepinfra/Gryphe/MythoMax-L2-13b

input $/1M

$0.08→$0.4↑
deepinfra/Gryphe/MythoMax-L2-13b

output $/1M

$0.09→$0.4↑
deepinfra/NousResearch/Hermes-3-Llama-3.1-70B

input $/1M

$0.3→$0.7↑
deepinfra/NousResearch/Hermes-3-Llama-3.1-70B

output $/1M

$0.3→$0.7↑
deepinfra/Qwen/Qwen2.5-72B-Instruct

input $/1M

$0.12→$0.36↑
deepinfra/Qwen/Qwen2.5-72B-Instruct

output $/1M

$0.39→$0.4↑
deepinfra/Qwen/Qwen3-14B

input $/1M

$0.06→$0.12↑
deepinfra/Qwen/Qwen3-235B-A22B-Instruct-2507

output $/1M

$0.6→$0.55↓
deepinfra/Qwen/Qwen3-30B-A3B

input $/1M

$0.08→$0.12↑
deepinfra/Qwen/Qwen3-30B-A3B

output $/1M

$0.29→$0.5↑
deepinfra/Qwen/Qwen3-32B

input $/1M

$0.1→$0.08↓
deepinfra/Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo

input $/1M

$0.29→$0.3↑
deepinfra/Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo

output $/1M

$1.2→$1↓
deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct

input $/1M

$0.14→$0.09↓
deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct

output $/1M

$1.4→$1.1↓
deepinfra/Sao10K/L3.1-70B-Euryale-v2.2

input $/1M

$0.65→$0.85↑
deepinfra/Sao10K/L3.1-70B-Euryale-v2.2

output $/1M

$0.75→$0.85↑
deepinfra/deepseek-ai/DeepSeek-V3

input $/1M

$0.38→$0.32↓
deepinfra/deepseek-ai/DeepSeek-V3-0324

input $/1M

$0.25→$0.24↓
deepinfra/deepseek-ai/DeepSeek-V3-0324

output $/1M

$0.88→$0.9↑
deepinfra/deepseek-ai/DeepSeek-V3.1

input $/1M

$0.27→$0.25↓
deepinfra/deepseek-ai/DeepSeek-V3.1

output $/1M

$1→$0.95↓
deepinfra/google/gemma-3-12b-it

output $/1M

$0.1→$0.15↑
deepinfra/google/gemma-3-27b-it

input $/1M

$0.09→$0.08↓
deepinfra/google/gemma-3-4b-it

input $/1M

$0.04→$0.05↑
deepinfra/google/gemma-3-4b-it

output $/1M

$0.08→$0.1↑
deepinfra/meta-llama/Llama-3.3-70B-Instruct-Turbo

input $/1M

$0.13→$0.1↓
deepinfra/meta-llama/Llama-3.3-70B-Instruct-Turbo

output $/1M

$0.39→$0.32↓
deepinfra/meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8

input $/1M

$0.15→$0.2↑
deepinfra/meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8

output $/1M

$0.6→$0.8↑
deepinfra/meta-llama/Llama-4-Scout-17B-16E-Instruct

input $/1M

$0.08→$0.1↑
deepinfra/meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo

input $/1M

$0.1→$0.4↑
deepinfra/meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo

output $/1M

$0.28→$0.4↑
deepinfra/meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo

output $/1M

$0.03→$0.04↑
deepinfra/mistralai/Mistral-Nemo-Instruct-2407

output $/1M

$0.04→$0.03↓
deepinfra/nvidia/NVIDIA-Nemotron-3.5-Lightning

input $/1M

$0.05→$0.08↑
deepinfra/openai/gpt-oss-120b

input $/1M

$0.05→$0.037↓
deepinfra/openai/gpt-oss-120b

output $/1M

$0.45→$0.17↓
deepinfra/openai/gpt-oss-20b

input $/1M

$0.04→$0.03↓
deepinfra/openai/gpt-oss-20b

output $/1M

$0.15→$0.14↓
gemini-2.5-flash-preview-09-2025

cache read $/1M

$0.075→$0.03↓
gemini-2.5-flash-lite-preview-06-17

cache read $/1M

$0.025→$0.01↓
gemini-2.5-pro-preview-tts

input $/1M

$1.25→$1↓
gemini-2.5-pro-preview-tts

output $/1M

$10→$20↑
gemini/gemini-2.5-flash-preview-09-2025

cache read $/1M

$0.075→$0.03↓
gemini/gemini-flash-latest

cache read $/1M

$0.075→$0.03↓
gemini/gemini-flash-lite-latest

cache read $/1M

$0.025→$0.01↓
gemini/gemini-2.5-flash-lite-preview-06-17

cache read $/1M

$0.025→$0.01↓
gemini/gemini-2.5-pro-preview-tts

input $/1M

$1.25→$1↓
gemini/gemini-2.5-pro-preview-tts

output $/1M

$10→$20↑
together_ai/Qwen/Qwen3.7-Max

input $/1M

$1.25→$2.5↑
together_ai/Qwen/Qwen3.7-Max

output $/1M

$3.75→$7.5↑
wandb/openai/gpt-oss-120b

input $/1M

$15000→$0.03↓
wandb/openai/gpt-oss-120b

output $/1M

$60000→$0.17↓
wandb/openai/gpt-oss-20b

input $/1M

$5000→$0.03↓
wandb/openai/gpt-oss-20b

output $/1M

$20000→$0.13↓
wandb/Qwen/Qwen3-Coder-480B-A35B-Instruct

input $/1M

$100000→$1↓
wandb/Qwen/Qwen3-Coder-480B-A35B-Instruct

output $/1M

$150000→$1.5↓
wandb/meta-llama/Llama-3.1-8B-Instruct

input $/1M

$22000→$0.22↓
wandb/meta-llama/Llama-3.1-8B-Instruct

output $/1M

$22000→$0.22↓
wandb/deepseek-ai/DeepSeek-V3.1

input $/1M

$55000→$0.55↓
wandb/deepseek-ai/DeepSeek-V3.1

output $/1M

$165000→$1.65↓
wandb/deepseek-ai/DeepSeek-V3.1

context window

128,000→161,000↑
wandb/meta-llama/Llama-3.3-70B-Instruct

input $/1M

$71000→$0.71↓
wandb/meta-llama/Llama-3.3-70B-Instruct

output $/1M

$71000→$0.71↓
novita/xiaomimimo/mimo-v2-flash

input $/1M

$0.1→$0.11↑
novita/xiaomimimo/mimo-v2-flash

output $/1M

$0.3→$0.33↑
novita/moonshotai/kimi-k2-thinking

max output tokens

262,144→100,352↓
novita/moonshotai/kimi-k2-0905

max output tokens

262,144→100,352↓
novita/qwen/qwen3-coder-480b-a35b-instruct

input $/1M

$0.3→$0.38↑
novita/qwen/qwen3-coder-480b-a35b-instruct

output $/1M

$1.3→$1.55↑
novita/moonshotai/kimi-k2-instruct

max output tokens

131,072→100,352↓
novita/deepseek/deepseek-v3-0324

max output tokens

163,840→65,536↓
novita/meta-llama/llama-3.3-70b-instruct

context window

131,072→12,288↓
novita/meta-llama/llama-3.3-70b-instruct

max output tokens

120,000→12,288↓
novita/qwen/qwen3-4b-fp8

max output tokens

20,000→8,192↓
novita/qwen/qwen2.5-7b-instruct

max output tokens

32,000→8,192↓
gemini-2.5-flash-native-audio-latest

input $/1M

$0.3→$0.5↑
gemini-2.5-flash-native-audio-latest

output $/1M

$2.5→$2↓
gemini-2.5-flash-native-audio-preview-09-2025

input $/1M

$0.3→$0.5↑
gemini-2.5-flash-native-audio-preview-09-2025

output $/1M

$2.5→$2↓
gemini-2.5-flash-native-audio-preview-12-2025

input $/1M

$0.3→$0.5↑
gemini-2.5-flash-native-audio-preview-12-2025

output $/1M

$2.5→$2↓
gemini/gemini-2.5-flash-native-audio-latest

input $/1M

$0.3→$0.5↑
gemini/gemini-2.5-flash-native-audio-latest

output $/1M

$2.5→$2↓
gemini/gemini-2.5-flash-native-audio-preview-09-2025

input $/1M

$0.3→$0.5↑
gemini/gemini-2.5-flash-native-audio-preview-09-2025

output $/1M

$2.5→$2↓
gemini/gemini-2.5-flash-native-audio-preview-12-2025

input $/1M

$0.3→$0.5↑
gemini/gemini-2.5-flash-native-audio-preview-12-2025

output $/1M

$2.5→$2↓
us.openai.gpt-5.6-sol

input $/1M

$5.5→$4.4↓
us.openai.gpt-5.6-sol

output $/1M

$33→$22↓
us.openai.gpt-5.6-sol

cache write $/1M

$6.875→$5.5↓
us.openai.gpt-5.6-sol

cache read $/1M

$0.55→$0.44↓
global.openai.gpt-5.6-sol

input $/1M

$5→$4↓
global.openai.gpt-5.6-sol

output $/1M

$30→$20↓
global.openai.gpt-5.6-sol

cache write $/1M

$6.25→$5↓
global.openai.gpt-5.6-sol

cache read $/1M

$0.5→$0.4↓
fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731

input $/1M

$0.14→$0.22↑
fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731

output $/1M

$0.28→$0.66↑
fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731

cache read $/1M

$0.028→$0.007↓

2026-08-26

together_ai/meta-llama/Llama-3.3-70B-Instruct-Turbo

input $/1M

$0.88→$1.04↑
together_ai/meta-llama/Llama-3.3-70B-Instruct-Turbo

output $/1M

$0.88→$1.04↑

2026-08-23

gpt-5.6

input $/1M

$5→$4↓
gpt-5.6

output $/1M

$30→$20↓
gpt-5.6

cache write $/1M

$6.25→$5↓
gpt-5.6

cache read $/1M

$0.5→$0.4↓
gpt-5.6-sol

input $/1M

$5→$4↓
gpt-5.6-sol

output $/1M

$30→$20↓
gpt-5.6-sol

cache write $/1M

$6.25→$5↓
gpt-5.6-sol

cache read $/1M

$0.5→$0.4↓
daybreak-blue-latest

input $/1M

$5→$4↓
daybreak-blue-latest

output $/1M

$30→$20↓
daybreak-blue-latest

cache write $/1M

$6.25→$5↓
daybreak-blue-latest

cache read $/1M

$0.5→$0.4↓
cognition/swe-1.7

input $/1M

$2.5→$0.5↓
cognition/swe-1.7

output $/1M

$12.5→$2.5↓
cognition/swe-1.7

cache read $/1M

$1→$0.2↓

2026-08-21

azure/gpt-5.6

context window

1,050,000→922,000↓
azure/gpt-5.6-sol

context window

1,050,000→922,000↓
azure/gpt-5.6-terra

context window

1,050,000→922,000↓
azure/gpt-5.6-luna

context window

1,050,000→922,000↓
azure/us/gpt-5.6

context window

1,050,000→922,000↓
azure/us/gpt-5.6-sol

context window

1,050,000→922,000↓
azure/us/gpt-5.6-terra

context window

1,050,000→922,000↓
azure/us/gpt-5.6-luna

context window

1,050,000→922,000↓
azure/eu/gpt-5.6

context window

1,050,000→922,000↓
azure/eu/gpt-5.6-sol

context window

1,050,000→922,000↓
azure/eu/gpt-5.6-terra

context window

1,050,000→922,000↓
azure/eu/gpt-5.6-luna

context window

1,050,000→922,000↓
claude-sonnet-4-6

max output tokens

64,000→128,000↑