Solar Pro 4

Fiche modèle · Upstage · solar-pro4

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.

Identité

OrganisationUpstage
Pays—
Famille—
Typellm
Date de sortie2026-08-06
Licenceproprietary
Licence commercialenon
Licence open sourcenon
Poids ouvertsnon
Auto-hébergement (DGX Spark)non — poids fermés
Modalitéstext
Fenêtre de contexte512 000
KV cache / token (Q8)—
Paramètres totaux—
Paramètres actifs (session)—
Source des paramètres actifs—
Curation des paramètres actifs—
Prix par session0.54 $ (1 000 000 tokens entrée + 200 000 sortie, prix le moins cher par fournisseur)
Tokens / tâche (Data Players)1.2 M
Tokens / tâche (Terminal-Bench)83.3 M (est.)
Ratio Terminal-Bench / Data Players69.4×
Coût / tâche (Terminal-Bench)4.54 $ (est.)
Tokens d'entraînement—
Coupure de connaissance—
Créé le2026-08-12T18:44:03.577717Z
Mis à jour le2026-11-04T06:05:12.974549Z
SourceLLM Stats
URLhttps://llm-stats.com/models/solar-pro4

Prix par fournisseur

Statut du modèle (offres + auto-hébergement) : non.

Toutes sources confondues : catalogue llm-stats, routeurs EUrouter/Cortecs et catalogues directs (OVHcloud, Scaleway, Infomaniak). Les prix sont convertis en USD pour la comparaison ; le prix publié et le taux BCE utilisé sont dans l'infobulle. Les prix inconnus restent indiqués « — ».

FournisseurSourceSouverainetéEntrée (/M tok)Sortie (/M tok)QuantificationContexteLatence (s)Débit (tok/s)Statut
Upstagellm-statsnon0.30 $1.20 $————active

Souveraineté qualifiée par offre (modèle × fournisseur) ; survolez le badge pour les critères.

Souveraineté

Scores par catégorie (top_scores)

CatégorieScore
language0.90
legal0.90
long_context0.70
math0.90
physics0.90
reasoning0.70
search0.50
finance0.90
frontend_development0.70
general0.70
healthcare0.90
agents0.40
biology0.90
chemistry0.90
code0.70
tool_calling0.40

Scores par benchmark

BenchmarkCatégorieScoreMaxNormalisé (API)VérifiéAuto-déclaréDaté du
aa-lcrreasoning0.711.000.71nonoui2026-11-04T06:05:12.980862Z
aime-2026math0.951.000.95nonoui2026-11-04T06:05:13.000238Z
browsecompreasoning0.491.000.49nonoui2026-11-04T06:05:13.003386Z
gpqareasoning0.891.000.89nonoui2026-11-04T06:05:12.987019Z
livecodebenchreasoning0.881.000.88nonoui2026-11-04T06:05:12.996737Z
mmlu-prolanguage0.861.000.86nonoui2026-11-04T06:05:12.993620Z
swe-bench-verifiedreasoning0.711.000.71nonoui2026-11-04T06:05:12.990265Z
tau3-bankingagents0.231.000.23nonoui2026-11-04T06:05:12.983892Z
terminal-bench-2.1reasoning0.571.000.57nonoui2026-11-04T06:05:12.977773Z

Classements

CatégorieRangScoreNote conservatriceBenchmarks évaluésMéthodeDaté duProvenance
agents123—11.403trueskill2026-10-19model_history
code104—20.793trueskill2026-10-19model_history

Lecture : Note conservatrice = borne basse TrueSkill (niveau estimé − marge d’incertitude) ; Benchmarks évalués = nombre de benchmarks de la catégorie pris en compte. À performances égales, un modèle évalué sur moins de benchmarks obtient une note plus basse : la note ne se déduit pas d’un seul score. Provenance : rankings = top 50 de /rankings ; model_history = historique du modèle (/models/{id}/history), utilisé pour les rangs au-delà du top 50.

Données

Dates de curation (récupération depuis l'API) : models 2026-11-04 · benchmarks 2026-11-04 · scores 2026-11-04 · updates 2026-11-04 · account 2026-11-04 · model_history 2026-11-04 · rankings 2026-11-04