Jaymr/new-api
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
60a1acb703a64186bf6eeef441e2fac947b75f26
new-api/setting/ratio_setting/model_ratio.go
T
Calcium-Ion 86ac0f7745 refactor: extract protocol conversion layer into standalone relaykit module (#6369) 
* test(relayconvert): add golden snapshot matrix and relaykit boundary guard

Phase 0 of the relaykit extraction plan: pin byte-level output of every
registered (from,to) request/response/stream conversion route, and
forbid kit-bound packages from growing host-only imports.

* wip(relayconvert): drop gin.Context from converter signatures; add convmeta draft

Phase 1 in progress: relayconvert now takes context.Context; host media
resolver adapts gin.Context back at the service boundary.

* refactor(relayconvert): decouple converters from RelayInfo, gin, and settings

Phase 1 of the relaykit extraction plan:
- converters now depend on convmeta.Meta (implemented by RelayInfo) instead
  of *relaycommon.RelayInfo; ClaudeConvertInfo and the format guesser move
  to convmeta with aliases left behind
- host settings reach converters via a convmeta.Options snapshot built in
  RelayInfo.ConvOptions; no more model_setting/reasoning global reads inside
  the conversion layer
- effort-suffix helpers move to service/relayconvert/reasoning (old package
  forwards); chat-to-responses upgrade policy moves to service (host routing
  logic, not conversion)
- golden conversion matrix unchanged

* test(relayconvert): tighten boundary — kit packages now free of gin/setting imports

* refactor(dto): drop gin and logger dependencies

Phase 2 (part 1): dto.Request.IsStream now takes *http.Request instead of
*gin.Context (Gemini's impl reads query/path off the std request); dto's
three logger calls become common.SysError. Boundary test allowlist is now
empty — kit-bound packages import no gin/setting/logger/model.

* refactor(kit): extract dependency-free kitutil; dto/types/relayconvert stop importing common

Phase 2 of the relaykit extraction plan:
- new service/relayconvert/kitutil holds the pure helpers the kit needs
  (JSON wrappers, pointer/string/uuid/timestamp utils, MaskSensitiveInfo,
  pluggable LogInfo/LogError hooks, Debug flag)
- dto, types, and all relayconvert packages now use kitutil; their only
  remaining internal deps are dto/types/constant
- common keeps every original symbol (MaskSensitiveInfo delegates to
  kitutil) so host code is untouched; main.go routes kit logging into
  common.SysLog/SysError and mirrors DebugEnabled
- golden conversion matrix unchanged

* refactor(kit): move EndpointType/FinishReason to types; OpenRouter dialect via Options

Kit packages (dto/types/relayconvert/reasonmap) no longer import constant:
- EndpointType and finish-reason values live in types; constant re-exports
- the OpenRouter special-case in claude->openai request conversion reads
  Options.OpenRouterDialect, set by the host from the channel type;
  InitChannelMeta invalidates the cached snapshot on channel switch

* refactor: extract relaykit submodule (dto/types/relayconvert/reasonmap)

Phase 3 of the relaykit extraction plan:
- new go module github.com/QuantumNous/new-api/relaykit containing dto
  (minus task family), types, relayconvert (with convmeta/kitutil/reasoning),
  and reasonmap; host consumes it via require + replace, go.work for dev
- task-family dto (task/suno/midjourney/video) stays in the host dto
  package; dual-consumer host files alias it as taskdto
- relaykit builds and tests standalone (GOWORK=off): no host imports,
  no gin, no DB, no settings
- golden conversion matrix unchanged

* build(docker): copy relaykit/go.mod before go mod download

The local-replace submodule's go.mod must exist inside the build context
for the main module graph to resolve.

* fix: address relaykit extraction regressions

* fix: address relaykit review regressions

* docs: document Meta nil receiver contract

* fix(relaykit): fail OpenAI→Claude conversion without max_tokens; reject negative default_max_tokens

The Claude Messages API requires max_tokens (omitting it is a 400
"Field required"), but with a nil Options.Claude.DefaultMaxTokens hook
the converters silently emitted a request the upstream is guaranteed to
reject. Both OpenAI Chat and Responses → Claude conversions now return
sharedclaude.ErrMissingMaxTokens when no path (client value, default
hook, thinking-adapter floor) supplied one. Unreachable in the host,
which always configures the hook.

Host side, claude.default_max_tokens now rejects negative values at the
option API before persisting — they would wrap into huge unsigned values
during conversion. Zero stays allowed: the current API treats
max_tokens: 0 as cache pre-warming.

* fix: make Gemini safety settings read path race-free
2026-07-27 15:56:21 +08:00

750 lines
27 KiB
Go
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
package ratio_setting
import (
"strings"
"github.com/QuantumNous/new-api/common"
"github.com/QuantumNous/new-api/relaykit/types"
"github.com/QuantumNous/new-api/setting/operation_setting"
)
// from songquanpeng/one-api
const (
USD2RMB = 7.3 // 暂定 1 USD = 7.3 RMB
USD = 500 // $0.002 = 1 -> $1 = 500
RMB = USD / USD2RMB
)
// modelRatio
// https://platform.openai.com/docs/models/model-endpoint-compatibility
// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Blfmc9dlf
// https://openai.com/pricing
// TODO: when a new api is enabled, check the pricing here
// 1 === $0.002 / 1K tokens
// 1 === ￥0.014 / 1k tokens
var defaultModelRatio = map[string]float64{
//"midjourney": 50,
"gpt-4-gizmo-*": 15,
"gpt-4o-gizmo-*": 2.5,
"gpt-4-all": 15,
"gpt-4o-all": 15,
"gpt-4": 15,
"gpt-4-0613": 15,
"gpt-4-32k": 30,
"gpt-4-32k-0613": 30,
"gpt-4-1106-preview": 5, // $10 / 1M tokens
"gpt-4-0125-preview": 5, // $10 / 1M tokens
"gpt-4-turbo-preview": 5, // $10 / 1M tokens
"gpt-4-vision-preview": 5, // $10 / 1M tokens
"gpt-4-1106-vision-preview": 5, // $10 / 1M tokens
"chatgpt-4o-latest": 2.5, // $5 / 1M tokens
"gpt-4o": 1.25, // $2.5 / 1M tokens
"gpt-4o-audio-preview": 1.25, // $2.5 / 1M tokens
"gpt-4o-audio-preview-2024-10-01": 1.25, // $2.5 / 1M tokens
"gpt-4o-2024-05-13": 2.5, // $5 / 1M tokens
"gpt-4o-2024-08-06": 1.25, // $2.5 / 1M tokens
"gpt-4o-2024-11-20": 1.25, // $2.5 / 1M tokens
"gpt-4o-realtime-preview": 2.5,
"gpt-4o-realtime-preview-2024-10-01": 2.5,
"gpt-4o-realtime-preview-2024-12-17": 2.5,
"gpt-4o-mini-realtime-preview": 0.3,
"gpt-4o-mini-realtime-preview-2024-12-17": 0.3,
"gpt-4.1": 1.0, // $2 / 1M tokens
"gpt-4.1-2025-04-14": 1.0, // $2 / 1M tokens
"gpt-4.1-mini": 0.2, // $0.4 / 1M tokens
"gpt-4.1-mini-2025-04-14": 0.2, // $0.4 / 1M tokens
"gpt-4.1-nano": 0.05, // $0.1 / 1M tokens
"gpt-4.1-nano-2025-04-14": 0.05, // $0.1 / 1M tokens
"gpt-image-1": 2.5, // $5 / 1M tokens
"o1": 7.5, // $15 / 1M tokens
"o1-2024-12-17": 7.5, // $15 / 1M tokens
"o1-preview": 7.5, // $15 / 1M tokens
"o1-preview-2024-09-12": 7.5, // $15 / 1M tokens
"o1-mini": 0.55, // $1.1 / 1M tokens
"o1-mini-2024-09-12": 0.55, // $1.1 / 1M tokens
"o1-pro": 75.0, // $150 / 1M tokens
"o1-pro-2025-03-19": 75.0, // $150 / 1M tokens
"o3-mini": 0.55,
"o3-mini-2025-01-31": 0.55,
"o3-mini-high": 0.55,
"o3-mini-2025-01-31-high": 0.55,
"o3-mini-low": 0.55,
"o3-mini-2025-01-31-low": 0.55,
"o3-mini-medium": 0.55,
"o3-mini-2025-01-31-medium": 0.55,
"o3": 1.0, // $2 / 1M tokens
"o3-2025-04-16": 1.0, // $2 / 1M tokens
"o3-pro": 10.0, // $20 / 1M tokens
"o3-pro-2025-06-10": 10.0, // $20 / 1M tokens
"o3-deep-research": 5.0, // $10 / 1M tokens
"o3-deep-research-2025-06-26": 5.0, // $10 / 1M tokens
"o4-mini": 0.55, // $1.1 / 1M tokens
"o4-mini-2025-04-16": 0.55, // $1.1 / 1M tokens
"o4-mini-deep-research": 1.0, // $2 / 1M tokens
"o4-mini-deep-research-2025-06-26": 1.0, // $2 / 1M tokens
"gpt-4o-mini": 0.075,
"gpt-4o-mini-2024-07-18": 0.075,
"gpt-4-turbo": 5, // $0.01 / 1K tokens
"gpt-4-turbo-2024-04-09": 5, // $0.01 / 1K tokens
"gpt-4.5-preview": 37.5,
"gpt-4.5-preview-2025-02-27": 37.5,
"gpt-5": 0.625,
"gpt-5-2025-08-07": 0.625,
"gpt-5-chat-latest": 0.625,
"gpt-5-mini": 0.125,
"gpt-5-mini-2025-08-07": 0.125,
"gpt-5-nano": 0.025,
"gpt-5-nano-2025-08-07": 0.025,
"gpt-5.5": 2.5, // $5 / 1M tokens
"gpt-5.6-sol": 2.5,
"gpt-5.6-terra": 1.25,
"gpt-5.6-luna": 0.5,
"gpt-3.5-turbo": 0.25,
"gpt-3.5-turbo-0613": 0.75,
"gpt-3.5-turbo-16k": 1.5, // $0.003 / 1K tokens
"gpt-3.5-turbo-16k-0613": 1.5,
"gpt-3.5-turbo-instruct": 0.75, // $0.0015 / 1K tokens
"gpt-3.5-turbo-1106": 0.5, // $0.001 / 1K tokens
"gpt-3.5-turbo-0125": 0.25,
"text-ada-001": 0.2,
"text-babbage-001": 0.25,
"text-curie-001": 1,
"text-davinci-edit-001": 10,
"code-davinci-edit-001": 10,
"whisper-1": 15, // $0.006 / minute -> $0.006 / 150 words -> $0.006 / 200 tokens -> $0.03 / 1k tokens
"tts-1": 7.5, // 1k characters -> $0.015
"tts-1-1106": 7.5, // 1k characters -> $0.015
"tts-1-hd": 15, // 1k characters -> $0.03
"tts-1-hd-1106": 15, // 1k characters -> $0.03
"davinci": 10,
"curie": 10,
"text-embedding-3-small": 0.01,
"text-embedding-3-large": 0.065,
"text-embedding-ada-002": 0.05,
"text-search-ada-doc-001": 10,
"text-moderation-stable": 0.1,
"text-moderation-latest": 0.1,
"claude-3-haiku-20240307": 0.125, // $0.25 / 1M tokens
"claude-3-5-haiku-20241022": 0.5, // $1 / 1M tokens
"claude-haiku-4-5-20251001": 0.5, // $1 / 1M tokens
"claude-3-sonnet-20240229": 1.5, // $3 / 1M tokens
"claude-3-5-sonnet-20240620": 1.5,
"claude-3-5-sonnet-20241022": 1.5,
"claude-3-7-sonnet-20250219": 1.5,
"claude-3-7-sonnet-20250219-thinking": 1.5,
"claude-sonnet-4-20250514": 1.5,
"claude-sonnet-4-5-20250929": 1.5,
"claude-opus-4-5-20251101": 2.5,
"claude-opus-4-6": 2.5,
"claude-opus-4-6-max": 2.5,
"claude-opus-4-6-high": 2.5,
"claude-opus-4-6-medium": 2.5,
"claude-opus-4-6-low": 2.5,
"claude-opus-4-7": 2.5,
"claude-opus-4-7-max": 2.5,
"claude-opus-4-7-xhigh": 2.5,
"claude-opus-4-7-high": 2.5,
"claude-opus-4-7-medium": 2.5,
"claude-opus-4-7-low": 2.5,
"claude-opus-4-8": 2.5,
"claude-opus-4-8-max": 2.5,
"claude-opus-4-8-xhigh": 2.5,
"claude-opus-4-8-high": 2.5,
"claude-opus-4-8-medium": 2.5,
"claude-opus-4-8-low": 2.5,
"claude-3-opus-20240229": 7.5, // $15 / 1M tokens
"claude-opus-4-20250514": 7.5,
"claude-opus-4-1-20250805": 7.5,
"ERNIE-4.0-8K": 0.120 * RMB,
"ERNIE-3.5-8K": 0.012 * RMB,
"ERNIE-3.5-8K-0205": 0.024 * RMB,
"ERNIE-3.5-8K-1222": 0.012 * RMB,
"ERNIE-Bot-8K": 0.024 * RMB,
"ERNIE-3.5-4K-0205": 0.012 * RMB,
"ERNIE-Speed-8K": 0.004 * RMB,
"ERNIE-Speed-128K": 0.004 * RMB,
"ERNIE-Lite-8K-0922": 0.008 * RMB,
"ERNIE-Lite-8K-0308": 0.003 * RMB,
"ERNIE-Tiny-8K": 0.001 * RMB,
"BLOOMZ-7B": 0.004 * RMB,
"Embedding-V1": 0.002 * RMB,
"bge-large-zh": 0.002 * RMB,
"bge-large-en": 0.002 * RMB,
"tao-8k": 0.002 * RMB,
"PaLM-2": 1,
"gemini-1.5-pro-latest": 1.25, // $3.5 / 1M tokens
"gemini-1.5-flash-latest": 0.075,
"gemini-2.0-flash": 0.05,
"gemini-2.5-pro-exp-03-25": 0.625,
"gemini-2.5-pro-preview-03-25": 0.625,
"gemini-2.5-pro": 0.625,
"gemini-2.5-flash-preview-04-17": 0.075,
"gemini-2.5-flash-preview-04-17-thinking": 0.075,
"gemini-2.5-flash-preview-04-17-nothinking": 0.075,
"gemini-2.5-flash-preview-05-20": 0.075,
"gemini-2.5-flash-preview-05-20-thinking": 0.075,
"gemini-2.5-flash-preview-05-20-nothinking": 0.075,
"gemini-2.5-flash-thinking-*": 0.075, // 用于为后续所有2.5 flash thinking budget 模型设置默认倍率
"gemini-2.5-pro-thinking-*": 0.625, // 用于为后续所有2.5 pro thinking budget 模型设置默认倍率
"gemini-2.5-flash-lite-preview-thinking-*": 0.05,
"gemini-2.5-flash-lite-preview-06-17": 0.05,
"gemini-2.5-flash": 0.15,
"gemini-robotics-er-1.5-preview": 0.15,
"gemini-embedding-001": 0.075,
"text-embedding-004": 0.001,
"chatglm_turbo": 0.3572, // ￥0.005 / 1k tokens
"chatglm_pro": 0.7143, // ￥0.01 / 1k tokens
"chatglm_std": 0.3572, // ￥0.005 / 1k tokens
"chatglm_lite": 0.1429, // ￥0.002 / 1k tokens
"glm-4": 7.143, // ￥0.1 / 1k tokens
"glm-4v": 0.05 * RMB, // ￥0.05 / 1k tokens
"glm-4-alltools": 0.1 * RMB, // ￥0.1 / 1k tokens
"glm-3-turbo": 0.3572,
"glm-4-plus": 0.05 * RMB,
"glm-4-0520": 0.1 * RMB,
"glm-4-air": 0.001 * RMB,
"glm-4-airx": 0.01 * RMB,
"glm-4-long": 0.001 * RMB,
"glm-4-flash": 0,
"glm-4v-plus": 0.01 * RMB,
"qwen-turbo": 0.8572, // ￥0.012 / 1k tokens
"qwen-plus": 10, // ￥0.14 / 1k tokens
"text-embedding-v1": 0.05, // ￥0.0007 / 1k tokens
"SparkDesk-v1.1": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v2.1": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v3.1": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v3.5": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v4.0": 1.2858,
"hunyuan": 7.143, // ¥0.1 / 1k tokens // https://cloud.tencent.com/document/product/1729/97731#e0e6be58-60c8-469f-bdeb-6c264ce3b4d0
// https://platform.lingyiwanwu.com/docs#-计费单元
// 已经按照 7.2 来换算美元价格
"yi-34b-chat-0205": 0.18,
"yi-34b-chat-200k": 0.864,
"yi-vl-plus": 0.432,
"yi-large": 20.0 / 1000 * RMB,
"yi-medium": 2.5 / 1000 * RMB,
"yi-vision": 6.0 / 1000 * RMB,
"yi-medium-200k": 12.0 / 1000 * RMB,
"yi-spark": 1.0 / 1000 * RMB,
"yi-large-rag": 25.0 / 1000 * RMB,
"yi-large-turbo": 12.0 / 1000 * RMB,
"yi-large-preview": 20.0 / 1000 * RMB,
"yi-large-rag-preview": 25.0 / 1000 * RMB,
"command": 0.5,
"command-nightly": 0.5,
"command-light": 0.5,
"command-light-nightly": 0.5,
"command-r": 0.25,
"command-r-plus": 1.5,
"command-r-08-2024": 0.075,
"command-r-plus-08-2024": 1.25,
"deepseek-chat": 0.27 / 2,
"deepseek-coder": 0.27 / 2,
"deepseek-reasoner": 0.55 / 2, // 0.55 / 1k tokens
// Perplexity online 模型对搜索额外收费，有需要应自行调整，此处不计入搜索费用
"llama-3-sonar-small-32k-chat": 0.2 / 1000 * USD,
"llama-3-sonar-small-32k-online": 0.2 / 1000 * USD,
"llama-3-sonar-large-32k-chat": 1 / 1000 * USD,
"llama-3-sonar-large-32k-online": 1 / 1000 * USD,
// grok
"grok-3-beta": 1.5,
"grok-3-mini-beta": 0.15,
"grok-2": 1,
"grok-2-vision": 1,
"grok-beta": 2.5,
"grok-vision-beta": 2.5,
"grok-3-fast-beta": 2.5,
"grok-3-mini-fast-beta": 0.3,
// submodel
"NousResearch/Hermes-4-405B-FP8": 0.8,
"Qwen/Qwen3-235B-A22B-Thinking-2507": 0.6,
"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": 0.8,
"Qwen/Qwen3-235B-A22B-Instruct-2507": 0.3,
"zai-org/GLM-4.5-FP8": 0.8,
"openai/gpt-oss-120b": 0.5,
"deepseek-ai/DeepSeek-R1-0528": 0.8,
"deepseek-ai/DeepSeek-R1": 0.8,
"deepseek-ai/DeepSeek-V3-0324": 0.8,
"deepseek-ai/DeepSeek-V3.1": 0.8,
}
var defaultModelPrice = map[string]float64{
"suno_music": 0.1,
"suno_lyrics": 0.01,
"dall-e-3": 0.04,
"imagen-3.0-generate-002": 0.03,
"black-forest-labs/flux-1.1-pro": 0.04,
"gpt-4-gizmo-*": 0.1,
"mj_video": 0.8,
"mj_imagine": 0.1,
"mj_edits": 0.1,
"mj_variation": 0.1,
"mj_reroll": 0.1,
"mj_blend": 0.1,
"mj_modal": 0.1,
"mj_zoom": 0.1,
"mj_shorten": 0.1,
"mj_high_variation": 0.1,
"mj_low_variation": 0.1,
"mj_pan": 0.1,
"mj_inpaint": 0,
"mj_custom_zoom": 0,
"mj_describe": 0.05,
"mj_upscale": 0.05,
"swap_face": 0.05,
"mj_upload": 0.05,
"sora-2": 0.3,
"sora-2-pro": 0.5,
"gpt-4o-mini-tts": 0.3,
"veo-3.0-generate-001": 0.4,
"veo-3.0-fast-generate-001": 0.15,
"veo-3.1-generate-preview": 0.4,
"veo-3.1-fast-generate-preview": 0.15,
}
var defaultAudioRatio = map[string]float64{
"gpt-4o-audio-preview": 16,
"gpt-4o-mini-audio-preview": 66.67,
"gpt-4o-realtime-preview": 8,
"gpt-4o-mini-realtime-preview": 16.67,
"gpt-4o-mini-tts": 25,
}
var defaultAudioCompletionRatio = map[string]float64{
"gpt-4o-realtime": 2,
"gpt-4o-mini-realtime": 2,
"gpt-4o-mini-tts": 1,
"tts-1": 0,
"tts-1-hd": 0,
"tts-1-1106": 0,
"tts-1-hd-1106": 0,
}
var modelPriceMap = types.NewRWMap[string, float64]()
var modelRatioMap = types.NewRWMap[string, float64]()
var completionRatioMap = types.NewRWMap[string, float64]()
var defaultCompletionRatio = map[string]float64{
"gpt-4-gizmo-*": 2,
"gpt-4o-gizmo-*": 3,
"gpt-4-all": 2,
"gpt-image-1": 8,
}
// InitRatioSettings initializes all model related settings maps
func InitRatioSettings() {
modelPriceMap.AddAll(defaultModelPrice)
modelRatioMap.AddAll(defaultModelRatio)
completionRatioMap.AddAll(defaultCompletionRatio)
cacheRatioMap.AddAll(defaultCacheRatio)
createCacheRatioMap.AddAll(defaultCreateCacheRatio)
imageRatioMap.AddAll(defaultImageRatio)
audioRatioMap.AddAll(defaultAudioRatio)
audioCompletionRatioMap.AddAll(defaultAudioCompletionRatio)
}
func GetModelPriceMap() map[string]float64 {
return modelPriceMap.ReadAll()
}
func ModelPrice2JSONString() string {
return modelPriceMap.MarshalJSONString()
}
func UpdateModelPriceByJSONString(jsonStr string) error {
return types.LoadFromJsonStringWithCallback(modelPriceMap, jsonStr, InvalidateExposedDataCache)
}
// GetModelPrice 返回模型的价格，如果模型不存在则返回-1，false
func GetModelPrice(name string, printErr bool) (float64, bool) {
name = FormatMatchingModelName(name)
if price, ok := modelPriceMap.Get(name); ok {
return price, true
}
if strings.HasSuffix(name, CompactModelSuffix) {
price, ok := modelPriceMap.Get(CompactWildcardModelKey)
if !ok {
if printErr {
common.SysError("model price not found: " + name)
}
return -1, false
}
return price, true
}
if printErr {
common.SysError("model price not found: " + name)
}
return -1, false
}
func UpdateModelRatioByJSONString(jsonStr string) error {
return types.LoadFromJsonStringWithCallback(modelRatioMap, jsonStr, InvalidateExposedDataCache)
}
// 处理带有思考预算的模型名称，方便统一定价
func handleThinkingBudgetModel(name, prefix, wildcard string) string {
if strings.HasPrefix(name, prefix) && strings.Contains(name, "-thinking-") {
return wildcard
}
return name
}
func GetModelRatio(name string) (float64, bool, string) {
name = FormatMatchingModelName(name)
ratio, ok := modelRatioMap.Get(name)
if !ok {
if strings.HasSuffix(name, CompactModelSuffix) {
if wildcardRatio, ok := modelRatioMap.Get(CompactWildcardModelKey); ok {
return wildcardRatio, true, name
}
//return 0, true, name
}
return 37.5, operation_setting.SelfUseModeEnabled, name
}
return ratio, true, name
}
func DefaultModelRatio2JSONString() string {
jsonBytes, err := common.Marshal(defaultModelRatio)
if err != nil {
common.SysError("error marshalling model ratio: " + err.Error())
}
return string(jsonBytes)
}
func GetDefaultModelRatioMap() map[string]float64 {
return defaultModelRatio
}
func GetDefaultModelPriceMap() map[string]float64 {
return defaultModelPrice
}
func CompletionRatio2JSONString() string {
return completionRatioMap.MarshalJSONString()
}
func UpdateCompletionRatioByJSONString(jsonStr string) error {
return types.LoadFromJsonStringWithCallback(completionRatioMap, jsonStr, InvalidateExposedDataCache)
}
func GetCompletionRatio(name string) float64 {
name = FormatMatchingModelName(name)
if strings.Contains(name, "/") {
if ratio, ok := completionRatioMap.Get(name); ok {
return ratio
}
}
hardCodedRatio, contain := getHardcodedCompletionModelRatio(name)
if contain {
return hardCodedRatio
}
if ratio, ok := completionRatioMap.Get(name); ok {
return ratio
}
return hardCodedRatio
}
type CompletionRatioInfo struct {
Ratio float64 `json:"ratio"`
Locked bool `json:"locked"`
}
func GetCompletionRatioInfo(name string) CompletionRatioInfo {
name = FormatMatchingModelName(name)
if strings.Contains(name, "/") {
if ratio, ok := completionRatioMap.Get(name); ok {
return CompletionRatioInfo{
Ratio: ratio,
Locked: false,
}
}
}
hardCodedRatio, locked := getHardcodedCompletionModelRatio(name)
if locked {
return CompletionRatioInfo{
Ratio: hardCodedRatio,
Locked: true,
}
}
if ratio, ok := completionRatioMap.Get(name); ok {
return CompletionRatioInfo{
Ratio: ratio,
Locked: false,
}
}
return CompletionRatioInfo{
Ratio: hardCodedRatio,
Locked: false,
}
}
func getHardcodedCompletionModelRatio(name string) (float64, bool) {
isReservedModel := strings.HasSuffix(name, "-all") || strings.HasSuffix(name, "-gizmo-*")
if isReservedModel {
return 2, false
}
if strings.HasPrefix(name, "gpt-") {
if strings.HasPrefix(name, "gpt-4o") {
if name == "gpt-4o-2024-05-13" {
return 3, true
}
if strings.HasPrefix(name, "gpt-4o-mini-tts") {
return 20, false
}
return 4, false
}
// gpt-5 匹配
if strings.HasPrefix(name, "gpt-5") {
if !strings.Contains(name, ".") {
return 8, true
}
if strings.HasPrefix(name, "gpt-5.4") {
if strings.HasPrefix(name, "gpt-5.4-nano") {
return 6.25, true
}
return 6, true
}
// gpt-5.5 and later models are unlocked
return 6, false
}
// gpt-4.5-preview匹配
if strings.HasPrefix(name, "gpt-4.5-preview") {
return 2, true
}
if strings.HasPrefix(name, "gpt-4-turbo") || strings.HasSuffix(name, "gpt-4-1106") || strings.HasSuffix(name, "gpt-4-1105") {
return 3, true
}
// 没有特殊标记的 gpt-4 模型默认倍率为 2
return 2, false
}
if strings.HasPrefix(name, "o1") || strings.HasPrefix(name, "o3") {
return 4, true
}
if name == "chatgpt-4o-latest" {
return 3, true
}
if strings.Contains(name, "claude-3") {
return 5, true
} else if strings.Contains(name, "claude-sonnet-4") || strings.Contains(name, "claude-opus-4") || strings.Contains(name, "claude-haiku-4") {
return 5, true
}
if strings.HasPrefix(name, "gpt-3.5") {
if name == "gpt-3.5-turbo" || strings.HasSuffix(name, "0125") {
// https://openai.com/blog/new-embedding-models-and-api-updates
// Updated GPT-3.5 Turbo model and lower pricing
return 3, true
}
if strings.HasSuffix(name, "1106") {
return 2, true
}
return 4.0 / 3.0, true
}
if strings.HasPrefix(name, "mistral-") {
return 3, true
}
if strings.HasPrefix(name, "gemini-") {
if strings.HasPrefix(name, "gemini-1.5") {
return 4, true
} else if strings.HasPrefix(name, "gemini-2.0") {
return 4, true
} else if strings.HasPrefix(name, "gemini-2.5-pro") { // 移除preview来增加兼容性，这里假设正式版的倍率和preview一致
return 8, false
} else if strings.HasPrefix(name, "gemini-2.5-flash") { // 处理不同的flash模型倍率
if strings.HasPrefix(name, "gemini-2.5-flash-preview") {
if strings.HasSuffix(name, "-nothinking") {
return 4, false
}
return 3.5 / 0.15, false
}
if strings.HasPrefix(name, "gemini-2.5-flash-lite") {
return 4, false
}
return 2.5 / 0.3, false
} else if strings.HasPrefix(name, "gemini-robotics-er-1.5") {
return 2.5 / 0.3, false
} else if strings.HasPrefix(name, "gemini-3-pro") {
if strings.HasPrefix(name, "gemini-3-pro-image") {
return 60, false
}
return 6, false
}
return 4, false
}
if strings.HasPrefix(name, "command") {
switch name {
case "command-r":
return 3, true
case "command-r-plus":
return 5, true
case "command-r-08-2024":
return 4, true
case "command-r-plus-08-2024":
return 4, true
default:
return 4, false
}
}
// hint 只给官方上4倍率，由于开源模型供应商自行定价，不对其进行补全倍率进行强制对齐
if strings.HasPrefix(name, "ERNIE-Speed-") {
return 2, true
} else if strings.HasPrefix(name, "ERNIE-Lite-") {
return 2, true
} else if strings.HasPrefix(name, "ERNIE-Character") {
return 2, true
} else if strings.HasPrefix(name, "ERNIE-Functions") {
return 2, true
}
switch name {
case "llama2-70b-4096":
return 0.8 / 0.64, true
case "llama3-8b-8192":
return 2, true
case "llama3-70b-8192":
return 0.79 / 0.59, true
}
return 1, false
}
func GetAudioRatio(name string) float64 {
name = FormatMatchingModelName(name)
if ratio, ok := audioRatioMap.Get(name); ok {
return ratio
}
return 1
}
func GetAudioCompletionRatio(name string) float64 {
name = FormatMatchingModelName(name)
if ratio, ok := audioCompletionRatioMap.Get(name); ok {
return ratio
}
return 1
}
func ContainsAudioRatio(name string) bool {
name = FormatMatchingModelName(name)
_, ok := audioRatioMap.Get(name)
return ok
}
func ContainsAudioCompletionRatio(name string) bool {
name = FormatMatchingModelName(name)
_, ok := audioCompletionRatioMap.Get(name)
return ok
}
func ModelRatio2JSONString() string {
return modelRatioMap.MarshalJSONString()
}
var defaultImageRatio = map[string]float64{
"gpt-image-1": 2,
}
var imageRatioMap = types.NewRWMap[string, float64]()
var audioRatioMap = types.NewRWMap[string, float64]()
var audioCompletionRatioMap = types.NewRWMap[string, float64]()
func ImageRatio2JSONString() string {
return imageRatioMap.MarshalJSONString()
}
func UpdateImageRatioByJSONString(jsonStr string) error {
return types.LoadFromJsonString(imageRatioMap, jsonStr)
}
func GetImageRatio(name string) (float64, bool) {
ratio, ok := imageRatioMap.Get(name)
if !ok {
return 1, false // Default to 1 if not found
}
return ratio, true
}
func AudioRatio2JSONString() string {
return audioRatioMap.MarshalJSONString()
}
func UpdateAudioRatioByJSONString(jsonStr string) error {
return types.LoadFromJsonStringWithCallback(audioRatioMap, jsonStr, InvalidateExposedDataCache)
}
func AudioCompletionRatio2JSONString() string {
return audioCompletionRatioMap.MarshalJSONString()
}
func UpdateAudioCompletionRatioByJSONString(jsonStr string) error {
return types.LoadFromJsonStringWithCallback(audioCompletionRatioMap, jsonStr, InvalidateExposedDataCache)
}
func GetModelRatioCopy() map[string]float64 {
return modelRatioMap.ReadAll()
}
func GetModelPriceCopy() map[string]float64 {
return modelPriceMap.ReadAll()
}
func GetCompletionRatioCopy() map[string]float64 {
return completionRatioMap.ReadAll()
}
func GetImageRatioCopy() map[string]float64 {
return imageRatioMap.ReadAll()
}
func GetAudioRatioCopy() map[string]float64 {
return audioRatioMap.ReadAll()
}
func GetAudioCompletionRatioCopy() map[string]float64 {
return audioCompletionRatioMap.ReadAll()
}
// 转换模型名，减少渠道必须配置各种带参数模型
func FormatMatchingModelName(name string) string {
if strings.HasPrefix(name, "gemini-2.5-flash-lite") {
name = handleThinkingBudgetModel(name, "gemini-2.5-flash-lite", "gemini-2.5-flash-lite-thinking-*")
} else if strings.HasPrefix(name, "gemini-2.5-flash") {
name = handleThinkingBudgetModel(name, "gemini-2.5-flash", "gemini-2.5-flash-thinking-*")
} else if strings.HasPrefix(name, "gemini-2.5-pro") {
name = handleThinkingBudgetModel(name, "gemini-2.5-pro", "gemini-2.5-pro-thinking-*")
}
if strings.HasPrefix(name, "gpt-4-gizmo") {
name = "gpt-4-gizmo-*"
}
if strings.HasPrefix(name, "gpt-4o-gizmo") {
name = "gpt-4o-gizmo-*"
}
return name
}
// result: 倍率or价格， usePrice， exist
func GetModelRatioOrPrice(model string) (float64, bool, bool) { // price or ratio
price, usePrice := GetModelPrice(model, false)
if usePrice {
return price, true, true
}
modelRatio, success, _ := GetModelRatio(model)
if success {
return modelRatio, false, true
}
return 37.5, false, false
}
Reference in New Issue View Git Blame Copy Permalink