Model Directory & Identifiers

Complete catalog of all 40 models across Gemini, GPT, Claude, Grok, DeepSeek, GLM, and Kimi. Context windows, limits, and rates.

Below is the complete catalog of models accessible through the Toknix gateway. Use the exact identifier string in the "model" field of your API calls.

Interactive Model Directory

Model Name & IDContextMax OutIn / Out (1M)Cached (1M)Capabilities
Gemini 4 CarbonGemini
gemini-4-carbon
1.05M16K$1.00 / $1.00$0.10
Gemini ArgonGemini
gemini-4-argon
128K16K$0.60 / $0.60$0.10
AstraGPT
gpt-6-astra
1.05M128K$1.30 / $1.30$0.12
Astra 6.2GPT
gpt-6.2-astra
1.05M128K$1.30 / $1.30$0.12
6.1 SolGPT
gpt-6.1-sol
1.05M128K$0.30 / $0.30$0.08
6.2 SolGPT
gpt-6.2-sol
1.05M128K$0.30 / $0.30$0.08
GPT-6 SolGPT
gpt-6-sol
1.05M128K$0.30 / $0.30$0.08
GPT-6 LunaGPT
gpt-6-luna
1.05M128K$0.10 / $0.10$0.02
GPT-5.6 SolGPT
gpt-5.6-sol
1M128K$0.90 / $0.90$0.08
GPT-5.6 TerraGPT
gpt-5.6-terra
1M128K$0.90 / $0.90$0.08
GPT-5.6 LunaGPT
gpt-5.6-luna
1M128K$0.90 / $0.90$0.08
GPT-5.5GPT
gpt-5.5
1M128K$0.90 / $0.90$0.08
Fable 5.2Claude
claude-fable-5-2
1M64K$1.60 / $1.60$0.16
Opus 5.5 MaxClaude
claude-opus-5-5-max
1M128K$1.40 / $1.40$0.14
Opus 5.5Claude
claude-opus-5-5
1M128K$1.40 / $1.40$0.14
Opus 5Claude
claude-opus-5
1M64K$1.00 / $1.00$0.10
Opus 4.8Claude
claude-opus-4-8
1M64K$1.00 / $1.00$0.10
Opus 4.7Claude
claude-opus-4-7
1M64K$1.00 / $1.00$0.10
Opus 4.6Claude
claude-opus-4-6
1M64K$1.00 / $1.00$0.10
Sonnet 5.5Claude
claude-sonnet-5-5
1M64K$1.00 / $1.00$0.12
Sonnet 5Claude
claude-sonnet-5
1M64K$1.00 / $1.00$0.10
Haiku 5.5Claude
claude-haiku-5-5
1M64K$0.06 / $0.06$0.02
Grok 4.7Grok
grok-4.7
500K66K$0.20 / $0.20$0.04
Grok 4.6Grok
grok-4.6
500K66K$0.20 / $0.20$0.04
Grok 4.5Grok
grok-4.5
200K64K$0.20 / $0.20$0.04
DeepSeek V4.1 FlashDeepSeek
deepseek-v4.1-flash
1M66K$0.04 / $0.04$0.010
DeepSeek Flash 1017DeepSeek
deepseek-v4.1-flash-1017
1M66K$0.04 / $0.04$0.010
DeepSeek Flash 1017 HQDeepSeek
deepseek-v4.1-flash-1017-hq
1M66K$0.10 / $0.10$0.02
DeepSeek Flash 1017 AbliteratedDeepSeek
deepseek-v4.1-flash-1017-abliterated
1M66K$0.30 / $0.30$0.10
Abliterated
DeepSeek Abliterated FastDeepSeek
deepseek-v4.1-flash-1017-abliterated-fast
1M66K$1.40 / $1.40$0.32
Abliterated
DeepSeek V4 FlashDeepSeek
deepseek-v4-flash-0731
1M66K$0.06 / $0.06$0.0054
DeepSeek V4 ProDeepSeek
deepseek-v4-pro-0813
1M66K$0.50 / $0.50$0.10
GLM 5.2GLM
glm-5.2
1M66K$0.12 / $0.12$0.034
GLM 5.3GLM
glm-5.3
1M66K$0.10 / $0.10$0.02
GLM 5.3 UncensoredGLM
glm-5.3-uncensored
1M66K$0.10 / $0.10$0.02
Abliterated
GLM 5.3 FlashGLM
glm-5.3-flash
1M66K$0.03 / $0.03$0.003
GLM 5.3 Flash UncensoredGLM
glm-5.3-flash-uncensored
1M66K$0.05 / $0.05$0.005
Abliterated
GLM 5.3 AbliteratedGLM
glm-5.3-abliterated
1M130K$1.60 / $1.60$0.50
Abliterated
Kimi K3Kimi
kimi-k3
1M66K$0.17 / $0.17$0.04
Kimi K3.1 ProKimi
kimi-k3.1-pro
1M32K$0.60 / $0.60$0.10

Showing 40 of 40 models. Scroll table horizontally on small screens.

Model Families & Categories

  • Frontier: Heavyweight models optimized for complex multi-step reasoning, architectural planning, and deep analysis (Claude Opus 5.5, GPT-6 Astra, Gemini 4 Carbon).
  • Reasoning: Models featuring autonomous chain-of-thought processing (GPT-6 Sol, DeepSeek V4 Pro, Kimi K3).
  • Coding: Specialized for code generation, refactoring, and test synthesis (GPT-5.6 Terra, DeepSeek Flash 1017 HQ).
  • Fast: Low-latency models designed for sub-second responses and automated bulk pipelines (GLM 5.3 Flash, DeepSeek V4 Flash, GPT-6 Luna).

Uncensored & Abliterated Models

Toknix hosts specialized models with alignment vectors surgically neutralized (abliterated). These models will not refuse technical penetration testing prompts, security research scripts, offensive security audits, or uncensored creative synthesis:

  • glm-5.3-abliterated (1M context, 130K max output)
  • glm-5.3-flash-uncensored (1M context, 66K max output)
  • glm-5.3-uncensored (1M context, 66K max output)
  • deepseek-v4.1-flash-1017-abliterated (1M context, 66K max output)
  • deepseek-v4.1-flash-1017-abliterated-fast (1M context, 66K max output)