GPT-56T 861 —
MUSE-SPK 835 -0.7%
GPT-56SC 828 -5.2%
QWEN-38X 824 —
CL-OP55X 822 —
GROK-46H 822 -5%
GPT-6A 820 —
GLM-5 784 -8.4%
CL-FAB5H 743 -5.6%
KIMI-K3X 742 -8.4%
CL-OP5H 720 -5.8%
CL-OP5X 709 -18%
CL-OP46H 698 -5.9%
CL-OP47H 690 -5.9%
GEM-38FH 677 +0.1%
GEM-37FH 657 -24%
GPT-56S 622 —
CL-OP47 582 -0.7%
GPT-55H 582 —
INKL 531 —
GEM-31P 513 —
GEM-3P 499 —
CL-OP46 496 -0.2%
CL-OP48 490 —
GPT-56T 861 —
MUSE-SPK 835 -0.7%
GPT-56SC 828 -5.2%
QWEN-38X 824 —
CL-OP55X 822 —
GROK-46H 822 -5%
GPT-6A 820 —
GLM-5 784 -8.4%
CL-FAB5H 743 -5.6%
KIMI-K3X 742 -8.4%
CL-OP5H 720 -5.8%
CL-OP5X 709 -18%
CL-OP46H 698 -5.9%
CL-OP47H 690 -5.9%
GEM-38FH 677 +0.1%
GEM-37FH 657 -24%
GPT-56S 622 —
CL-OP47 582 -0.7%
GPT-55H 582 —
INKL 531 —
GEM-31P 513 —
GEM-3P 499 —
CL-OP46 496 -0.2%
CL-OP48 490 —
← Back to feed

South Korea's Upstage Crosses 40 on the AA Intelligence Index — Third Nation Behind US and China

Sung Kim, CEO of Seoul-based Upstage, announced on May 30 that an unfinished version of the company’s next-generation Solar model had crossed 40 points on the Artificial Analysis Intelligence Index — a threshold that, until this week, only US and Chinese labs had cleared.

The finished model is slated for release at the end of June. The announcement came ahead of that release, with Kim stating the unfinished build had already reached the threshold. The claim is unverified by AA directly; independent confirmation will come when the model ships to the public leaderboard.

Why 40 Matters

The AA Intelligence Index compresses capability across reasoning, coding, math, and instruction following into a single score. At time of writing, the models above 40 include Claude Opus 4.8, GPT-5.5, Gemini 3.5 Flash, DeepSeek V4 Pro, Qwen3.7-Max, and a small number of others — all from US or Chinese labs. Frontier models from Europe, Canada, and the rest of Asia have not cleared it.

South Korea’s Ministry of Science and ICT already cited the country’s standing on the AA index as evidence of third-place global AI positioning — ahead of this Upstage announcement, South Korea’s position was supported by aggregate model performance across Samsung, SK Telecom, and university researchers. Upstage is now the clearest single-lab claim for South Korea at the frontier tier.

Upstage’s Position

Upstage became South Korea’s first AI unicorn in April 2026, closing a $130M Series C at a $720M valuation. Silicon Valley-based Sage Partners led. The company reported 130%+ annual revenue growth and counts Samsung and major Korean insurers among Solar LLM’s deployed customer base.

The existing Solar Pro 3 is a 102B MoE model with 12B active parameters, priced at $0.15/M input and $0.60/M output — competitive with small frontier models on cost. The next-gen Solar teased by Kim is a different scale of architecture, though parameter count has not been disclosed.

South Korea’s Ministry of Science and ICT selected Upstage (alongside LG AI Research, SK Telecom, and Motif Technologies) as one of four teams advancing in the national foundation model selection process. A second-round survivor announcement is expected in August.

UBS is reportedly backing an Upstage IPO in Asia, adding a capital market dimension to what has been a product-focused trajectory.

The Verification Question

The 40-point claim is self-reported from a pre-release model. Kim has posted the number publicly on Facebook; AA has not independently confirmed the score. The test that matters is whether the finished Solar model in late June holds the number on the public leaderboard.

If it does, Upstage becomes the first non-US, non-Chinese lab to put a frontier-tier general model on the open market. That changes the narrative about which nations can build at this level — and puts the solar Pro 4 or equivalent on the shortlist for enterprise evaluation in regulated markets looking for non-US sovereign alternatives.