Will there be support for more Models?

[YuJian920]

For example, supporting Gemini 2.0 Pro/Flash Think, o1, o3-mini, etc., providing support based on OpenRouter doesn't seem difficult.

[mrcfps]

You can customize models to use via inserting new records into refly.model_infos table in refly_db Postgres instance, here is an example:

INSERT INTO "refly"."model_infos" ("name", "label", "provider", "tier", "enabled", "updated_at", "context_limit", "max_output", "capabilities")
VALUES ('deepseek/deepseek-r1', 'DeepSeek R1', 'deepseek', 't1', 't', now(), 64000, 8192, '{}');

[YuJian920]

You can customize models to use via inserting new records into refly.model_infos table in refly_db Postgres instance, here is an example:你可以通过在 refly_db Postgres 实例的 refly.model_infos 表中插入新记录来自定义要使用的模型,这是一个例子:

INSERT INTO "refly"."model_infos" ("name", "label", "provider", "tier", "enabled", "updated_at", "context_limit", "max_output", "capabilities")
VALUES ('deepseek/deepseek-r1', 'DeepSeek R1', 'deepseek', 't1', 't', now(), 64000, 8192, '{}');

In fact, what I actually wanted to ask is whether your paid service will update with new models😭

[mrcfps]

Hi, sorry for the late response. We just added o3-mini model to our hosted service!

For the o1 model, the token pricing is so hard for us to maintain a balanced budget at this time😭.

The Gemini 2.0 Pro/Flash Thinking is still experimental and suffers from poor uptime. We'll add these models when they have reached GA.

[YuJian920]

Hi, sorry for the late response. We just added o3-mini model to our hosted service!

For the o1 model, the token pricing is so hard for us to maintain a balanced budget at this time😭.

The Gemini 2.0 Pro/Flash Thinking is still experimental and suffers from poor uptime. We'll add these models when they have reached GA.

Thanks for the reply!