跳转到主要内容

Solar Pro 4

UpstageProprietary

描述

Solar Pro 4 is Upstage's proprietary reasoning-focused agentic LLM (text→text) for finishing multi-step commercial work: long documents, terminal tasks, and multi-turn tool use. Supports English, Korean, and Japanese with a ~512K context window (OpenRouter listing ~524K) and up to 128K output tokens. Reasoning effort is adjustable (high for deep analysis, low for fast chat). Vendor/AA self-reported highlights include Terminal-Bench v2.1 57, τ³-Banking 23, AA-LCR 71, and GPQA Diamond 89.0; in-house asterisk rows include BrowseComp 49.2, SWE-Bench Verified (OpenHands) 70.6, MMLU-Pro 86.3, LiveCodeBench 87.8, AIME 2026 95.3. GDPval-AA v2 38.8 is noted here only — catalog gdpval-aa is Elo-scale (max_score 3000) and 38.8 does not fit that encoding (no gdpval-aa-v2 id). Standard list pricing is $0.30 input / $0.06 cached input / $1.20 output per 1M tokens; launch promo is 90% off on Upstage Console and OpenRouter through September 10, 2026 (23:59 UTC) — promo not stored in the provider price row. Also available via SolarChat, Hermes Agent, and Upstage Studio.

发布日期
2026-08-06
参数规模
—
上下文长度
524K
支持模态
text

能力雷达图

28
general
51
coding
89
reasoning
62
science
40
agents
0
multimodal

排行榜排名

领域#排名分数来源
智能体能力模型榜59
47.0
LS
代码能力榜167
71.0
AA
通用能力榜266
46.0
AA
科学能力142
67.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

Tau3 Banking23.0%自报

Language

MMLU-Pro86.3%自报

Math

AIME 202695.3%自报

Reasoning

GPQANYU + Cohere + Anthropic (2023)89.0%自报
LiveCodeBench87.8%自报
AA-LCR71.0%自报
SWE-Bench Verified70.6%自报
Terminal-Bench 2.157.0%自报
BrowseCompOpenAI (2025)49.2%自报

AA 评测指数

(Artificial Analysis)
Gpqa(NYU + Cohere + Anthropic (2023))
89.1
Lcr(Artificial Analysis)
74.0
Terminalbench V2 1
57.3
Coding Index(Artificial Analysis)
52.7
Scicode(UIUC + Argonne National Lab (2024))
44.6
Hle(Center for AI Safety + Scale AI (2025))
29.2
Intelligence Index(Artificial Analysis)
28.2
Tau Banking
23.3
Terminalbench V4 0
0.5

LLM Stats 分类评分

(LLM Stats (zeroeval))
Language
90
Legal
90
Math
90
Physics
90
Finance
90
Healthcare
90
Biology
90
Chemistry
90
Long Context
70
Reasoning
70
Frontend Development
70
General
70
Code
70
Search
50
Agents
40
Tool Calling
40

定价

输入价格$0.3 / 1M tokens
输出价格$1.2 / 1M tokens
混合价格(3:1)$0.525 / 1M tokens
缓存读取价格$0.06 / 1M tokens

速度

Tokens/秒93.0
首Token延迟1.35s
首回答延迟22.86s

供应商价格排行

供应商价格排行

5 个供应商

最便宜: Upstage最贵: Kilo Gateway
供应商输入输出
1Upstage最便宜
$0
$0
2NanoGPT
$0.03
$0.12
3LLMTR
$0.03
$0.12
4OpenRouter
$0.09
$0.36
5Kilo Gateway
$0.3
$1.2

比较该模型在不同 API 供应商之间的定价。

外部链接