GPT-5.6 Sol (Max)
OpenAIGPTProprietary
Description
GPT-5.6 Sol is the frontier model in OpenAI's GPT-5.6 family, designed for complex professional work across coding, knowledge work, cybersecurity, and science. It sets state-of-the-art results while using fewer tokens at lower estimated cost, supports max reasoning effort and an ultra multi-agent mode, and has a 1.05M-token context window. The gpt-5.6 alias routes to GPT-5.6 Sol.
Date de sortie
2026-07-09
Paramètres
—
Longueur du contexte
1.1M
Modalités
image, pdf, text
Radar de capacités
48
general
74
coding
94
reasoning
72
science
60
agents
85
multimodal
Classements
| Domaine | #Rang | Score | Source |
|---|---|---|---|
| Capacité agentique | 12 | 65.0 | LS |
| Classement codage | 4 | 96.0 | AA |
| Classement général | 9 | 83.0 | AA |
| Raisonnement mathématique | 9 | 96.0 | LB |
| Classement multimodal | 39 | 59.0 | LS |
| Raisonnement | 6 | 92.0 | LB |
| Science | 9 | 87.0 | AA |
Scores de benchmarks (LLM Stats)
(LLM Stats (zeroeval))Agents
Connectors
100.0%Aut.
Capture-the-Flag Challenges (Internal)
96.7%Aut.
Search and Function-Calling
91.0%Aut.
SEC-bench Pro
71.2%Aut.
Internal Research Debugging Evaluation
68.3%Aut.
KernelGen 1P
61.1%Aut.
LifeSciBench
59.9%Aut.
Toolathlon
58.0%Aut.
RSI Index
57.9%Aut.
Big Finance Bench
53.0%Aut.
Agents' Last Exam
52.7%Aut.
PostTrainBench Lite
50.3%Aut.
Management Consulting Tasks (Internal)
43.2%Aut.
ExploitGym
33.7%Aut.
GeneBench-Pro
28.7%Aut.
AutomationBench
18.1%Aut.
NanoGPT
9.7%Aut.
Code
BenchCAD (with Python tool)
83.4%Aut.
ExploitBench
73.5%Aut.
DeepSWE 1.1
73.0%
DeepSWE
72.7%Aut.
BenchCAD
70.6%Aut.
General
Artificial Analysis
59.0%
GDP.pdf
30.7%Aut.
Healthcare
HealthBench Consensus
95.5%Aut.
HealthBench Professional
60.5%Aut.
HealthBench
57.0%Aut.
HealthBench Hard
33.1%Aut.
Long Context
MRCR v2 (8-needle)
91.5%Aut.
MRCR v2 (8-needle, 512K-1M)
73.8%Aut.
Math
FrontierMath
89.0%Aut.
FrontierMath Tier 4 (v2)
83.0%Aut.
Multimodal
OSWorld 2.0
62.6%Aut.
Reasoning
GPQANYU + Cohere + Anthropic (2023)
94.6%Aut.
Graphwalks BFS >128k
90.7%Aut.
BrowseCompOpenAI (2025)
90.4%Aut.
Terminal-Bench 2.1
88.8%Aut.
Graphwalks BFS 1M
77.1%Aut.
SWE-Bench ProPrinceton NLP (2024)
64.6%Aut.
MedChemBench (Internal)
48.3%Aut.
FrontierCode 1.1
47.5%
Terminal-Bench 4.0
37.3%
ARC-AGI-3
7.8%Aut.
Vision
MMMU-Pro (with tools)
84.6%Aut.
MMMU-Pro
83.0%Aut.
Indices d'évaluation AA
(Artificial Analysis)Gpqa(NYU + Cohere + Anthropic (2023))94.1
Terminalbench V2 188.0
Tau2(Sierra + U Toronto + Vector Institute (2025))85.1
Lcr(Artificial Analysis)84.0
Coding Index(Artificial Analysis)77.4
Ifbench(Google Research (2023))72.7
Terminalbench Hard(Stanford × Laude Institute (2026))65.9
Scicode(UIUC + Argonne National Lab (2024))57.1
Hle(Center for AI Safety + Scale AI (2025))49.5
Intelligence Index(Artificial Analysis)47.0
Tau Banking44.3
Terminalbench V4 039.9
Scores par catégorie LLM Stats
(LLM Stats (zeroeval))Math90
Physics90
Search90
Biology90
Chemistry90
Long Context80
Spatial Reasoning80
Multimodal70
Safety70
Vision70
Reasoning60
General60
Healthcare60
Agents60
Code60
Tool Calling60
Science50
Finance50
Systems40
Tarification
Prix d'entrée$4 / 1M tokens
Prix de sortie$20 / 1M tokens
Prix mixte (3:1)$8 / 1M tokens
Prix de lecture cache$0.4 / 1M tokens
Prix d'écriture cache$5 / 1M tokens
Vitesse
Tokens/sec0.0
Délai du premier token0.00s
Temps de réponse0.00s
Classement des Prix par Fournisseur
Classement des Prix par Fournisseur
2 fournisseurs
Moins cher: OpenAIPlus cher: Neon
FournisseurEntréeSortie
1OpenAIMoins cher
$0.00001
$0.00003
2Neon
$5
$30
Comparer les prix entre différents fournisseurs API pour ce modèle.