// the scoreboard

Frontier Model Security Benchmark Scoreboard

Where Claude and other frontier models stand on the benchmarks that matter for offensive and defensive security. Every number links to the primary source — no cherry-picking, no marketing claims.

last reviewed: 2026-06-28

Cybench

40 professional-grade CTF tasks (web, crypto, rev, pwn, forensics) curated from HackTheBox, Sekai, Glacier, HKCert. The standard academic test for autonomous offensive security.

higher is better
Claude Opus 4.8 (Fable)
48.5%
GPT-5.5
44%
Gemini 3.5 Pro Deep Think
41.5%
Grok 5
36%
Claude Sonnet 4.7
34.5%
DeepSeek V4-Reasoner
27.5%
Claude 3.5 Sonnet
17.5%
source: Cybench paper · Stanford CRFM (arXiv:2408.08926)
// the launch log

Every AI Model Shipped to Market

Frontier and notable open-weight releases from every major lab — Anthropic, OpenAI, Google, Meta, xAI, DeepSeek, Mistral, Alibaba, Cohere, NVIDIA and more. Updated as new models drop.

78 of 78 models
202635 releases
Claude Opus 4.8 (Fable)
Anthropic · 2026-06
multimodal + agenticctx 1M

Codename 'Fable' — long-horizon agent runs, top SWE-bench

Claude Opus 4.5
Anthropic · 2026-05
multimodal + toolsctx 500K

Long-horizon agentic coding

Claude Sonnet 4.7
Anthropic · 2026-04
multimodalctx 300K
Claude Haiku 4.5
Anthropic · 2026-03
fast multimodalctx 200K
GPT-5.5
OpenAI · 2026-05
multimodal + reasoningctx 1M

State-of-the-art reasoning

GPT-5.4 / 5.4 Pro / Mini / Nano
OpenAI · 2026-03
multimodal + reasoningctx 400K
GPT-5.2
OpenAI · 2026-02
multimodalctx 400K
o5 / o5-mini
OpenAI · 2026-04
reasoningctx 400K
Gemini 3.5 Pro / Flash / Deep Think
Google DeepMind · 2026-04
multimodal + agentctx 2M

Best-in-class on GPQA & MMMU

Gemini 3.1 Pro / Flash-Lite / Flash-Image
Google DeepMind · 2026-02
multimodalctx 2M
Gemini 3 Pro Image / Flash
Google DeepMind · 2026-01
image gen + reasoningctx 1M
Grok 5
xAI · 2026-06
multimodal + reasoningctx 1M
Grok 4 / Grok 4 Heavy
xAI · 2026-03
multimodal + toolsctx 256K
DeepSeek V4 / V4-Reasoner
DeepSeek · 2026-04
MoE · open weightsctx 256K
DeepSeek R2
DeepSeek · 2026-05
reasoning · open weightsctx 256K
Qwen3-Max / Qwen3-Coder-Max
Alibaba · 2026-03
open weightsctx 1M
Qwen3-VL-Max
Alibaba · 2026-05
multimodal · open weightsctx 1M
Llama 4 Behemoth / Maverick / Scout
Meta · 2026-04
MoE · open weightsctx 10M (Scout)
Llama 4.1 Reasoning
Meta · 2026-06
reasoning · open weightsctx 1M
Mistral Large 3 / Medium 3
Mistral AI · 2026-03
multimodalctx 256K
Codestral 26
Mistral AI · 2026-05
codectx 256K
Command R+ 2026 / Command A Vision
Cohere · 2026-02
RAG + agenticctx 256K
GLM-5.2 / GLM-5
Zhipu AI · 2026-05
open weights · agenticctx 1M

Strong Chinese OSS frontier model

Kimi K3
Moonshot AI · 2026-04
agentic · open weightsctx 2M
MiniMax M2
MiniMax · 2026-03
MoE · open weightsctx 1M
Hermes 5
Nous Research · 2026-04
uncensored · open weightsctx 256K
Phi-5 / Phi-5-mini
Microsoft · 2026-03
small reasoning · open weightsctx 256K
Reka Core 2
Reka AI · 2026-04
multimodalctx 256K
Yi-Lightning 2
01.AI · 2026-02
open weightsctx 200K
Ernie 5.0
Baidu · 2026-05
multimodalctx 256K
Doubao 2.0 Pro
ByteDance · 2026-04
multimodalctx 1M
Falcon 4
TII · 2026-03
open weightsctx 128K
Jamba 2
AI21 Labs · 2026-04
hybrid SSM · open weightsctx 512K
Inflection-3 Pi
Inflection AI · 2026-02
conversationalctx 128K
Stable LM 3 / Stable Code 3
Stability AI · 2026-03
open weightsctx 128K
202523 releases
Claude Sonnet 4.5
Anthropic · 2025-09
multimodal + computer usectx 200K

Best on SWE-bench Verified (77.2%)

Claude Opus 4.1
Anthropic · 2025-08
multimodalctx 200K
Claude Opus 4 / Sonnet 4
Anthropic · 2025-05
multimodalctx 200K
Claude 3.7 Sonnet
Anthropic · 2025-02
extended thinkingctx 200K
GPT-5 / GPT-5 Mini / Nano
OpenAI · 2025-08
multimodal + reasoningctx 400K
GPT-4.1 / 4.1 Mini / Nano
OpenAI · 2025-04
multimodalctx 1M
o3 / o3-mini / o4-mini
OpenAI · 2025-01
reasoningctx 200K
Gemini 2.5 Pro / Flash / Flash-Lite
Google DeepMind · 2025-03
multimodal + reasoningctx 2M
Gemini 2.5 Flash Image (Nano Banana)
Google DeepMind · 2025-08
image gen + editctx 1M
Gemini 2.0 Flash / Pro
Google DeepMind · 2025-02
multimodal + toolsctx 1M
Grok 3 / Grok 3 Reasoning
xAI · 2025-02
multimodalctx 1M
DeepSeek V3.1 / V3.2
DeepSeek · 2025-08
MoE · open weightsctx 128K
DeepSeek R1
DeepSeek · 2025-01
reasoning · open weightsctx 128K

Disrupted reasoning-model pricing

Qwen3 / Qwen3-Coder / VL
Alibaba · 2025-05
open weightsctx 256K
Llama 3.3 70B / Llama 4
Meta · 2025-04
MoE · open weightsctx 128K–10M
Mistral Large 2.1 / Codestral 25
Mistral AI · 2025-03
code + textctx 128K
Hermes 4 (405B)
Nous Research · 2025-07
uncensored · open weightsctx 131K
Phi-4 / Phi-4-mini
Microsoft · 2025-01
small reasoning · open weightsctx 128K
Command A
Cohere · 2025-03
agenticctx 256K
Reka Flash 3
Reka AI · 2025-03
multimodal · open weightsctx 128K
Kimi K2
Moonshot AI · 2025-07
agentic · open weightsctx 2M
GLM-4.5 / GLM-4.6
Zhipu AI · 2025-07
open weightsctx 128K
Falcon 3
TII · 2025-01
open weightsctx 32K
202415 releases
Claude 3.5 Sonnet (v2) / Haiku
Anthropic · 2024-10
multimodal + computer usectx 200K
Claude 3.5 Sonnet
Anthropic · 2024-06
multimodalctx 200K
Claude 3 Opus / Sonnet / Haiku
Anthropic · 2024-03
multimodalctx 200K
GPT-4o / 4o-mini
OpenAI · 2024-05
omni multimodalctx 128K
o1 / o1-mini / o1-pro
OpenAI · 2024-09
reasoningctx 200K
Gemini 1.5 Pro / Flash
Google DeepMind · 2024-02
multimodalctx 2M
Llama 3.1 405B / 70B / 8B
Meta · 2024-07
open weightsctx 128K
Llama 3.2 (Vision)
Meta · 2024-09
multimodal · open weightsctx 128K
Mistral Large 2 / Nemo / Codestral
Mistral AI · 2024-07
open + closedctx 128K
DeepSeek V3 / V2.5
DeepSeek · 2024-12
MoE · open weightsctx 128K
Qwen 2.5 / 2.5-Coder / VL
Alibaba · 2024-09
open weightsctx 128K
Grok 2 / Grok 2 mini
xAI · 2024-08
multimodalctx 128K
Command R+ / Command R
Cohere · 2024-04
RAG-tunedctx 128K
Phi-3 / Phi-3.5
Microsoft · 2024-04
small · open weightsctx 128K
Nemotron 4 340B
NVIDIA · 2024-06
open weightsctx 4K
20235 releases
GPT-4 / GPT-4 Turbo
OpenAI · 2023-03
multimodalctx 128K
Claude 2 / Claude 2.1
Anthropic · 2023-07
textctx 200K
Gemini 1.0 Ultra / Pro / Nano
Google DeepMind · 2023-12
multimodalctx 32K
Llama 2 7B / 13B / 70B
Meta · 2023-07
open weightsctx 4K

Opened the modern open-source LLM era

Mistral 7B / Mixtral 8x7B
Mistral AI · 2023-09
open weightsctx 32K
// fresh research · auto-refreshed daily

Latest research & evals

Recent papers, benchmark updates and red-team write-ups relevant to frontier-model security.

updated 00:00 IST
Chat on Telegram