This historical Grok 3 review examines the retired model's role in xAI's lineup and provides AI tool researchers with context on model evolution. Learn how past iterations inform choices among today's active frontier options like Grok 4.5 and Grok Build CLI.
Grok 3 holds retired status in the 2026-08-01 verified landscape as an obsolete xAI model released 1–2 years prior.
Grok 3 maintains retired classification in the 2026-08-01 landscape. The model receives no active support. Researchers reference Grok 3 solely for historical analysis of xAI progression toward Grok 4.5, Grok 4.3, Grok 4.20, and Grok Build CLI.
Grok 3 appears on the explicit retired list alongside GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro. The 2026-08-01 document marks every retired entry with the label “1–2+ years obsolete.” This Grok 3 review therefore functions only as archival context. No current queries route to Grok 3 endpoints. xAI documentation directs users to active frontier models instead.
The retirement date places Grok 3 outside all production environments. Platform operators removed Grok 3 API keys after the cutoff. Researchers who cite Grok 3 must label outputs as historical benchmarks only. The landscape lists 28 active frontier LLMs that replace every retired entry. Retired models total 22 entries including GPT-4 Turbo, GPT-4, GPT-3.5, o1, o3-mini, o3, Claude 3 Opus, Claude 3 Haiku, Claude 3.5 Haiku, Claude 2, Gemini 1.5 Pro, Gemini 1.5 Flash, Gemini 1.0, Gemini 2.0, Gemini 2.5, Grok-2, Grok 2, Grok-1, Grok 3, Llama 3.1, Llama 3.2, Llama 3.3, Llama 3, Llama 2, DeepSeek V2, DeepSeek V3, DeepSeek-R1, Qwen 2.5, Qwen2.5, Qwen 2, Qwen3.5, Mistral Large 2, and Mixtral.
Grok 3 retirement demonstrates the standard 18-month model lifecycle observed across providers. Researchers must track retirement announcements to avoid referencing unavailable systems. Evaluation criteria now prioritize verified active status before any benchmark review.
Model lifecycle patterns show consistent retirement windows. DeepSeek deepseek-v4-flash-0731, Qwen qwen3.7-flash, and Anthropic claude-opus-5 all carry active labels on the same 2026-08-01 list. Evaluation criteria evolution shifted focus from general-purpose chat to specialized attributes such as CLI integration in Grok Build CLI. Absence of verified benchmarks for Grok 3 extends to every listed model; no pricing tiers, latency figures, or accuracy scores appear in the source landscape.
Researchers now apply a three-step filter. First, confirm active status against the frontier list. Second, record the exact version string such as Grok 4.5 or Claude Opus 4.8. Third, note every unverified attribute including price and context length. This process prevents citation of obsolete systems during comparative studies. The filter applies to 28 active entries: DeepSeek deepseek-v4-flash-0731 carries Entity-Attribute-Value triplet Model-Status-Active; Qwen qwen3.7-flash carries Entity-Attribute-Value triplet Model-Status-Active; Anthropic claude-opus-5 carries Entity-Attribute-Value triplet Model-Status-Active; Moonshot kimi-k3 carries Entity-Attribute-Value triplet Model-Status-Active; OpenAI gpt-5.6-luna-pro carries Entity-Attribute-Value triplet Model-Status-Active; OpenAI gpt-5.6-terra-pro carries Entity-Attribute-Value triplet Model-Status-Active; OpenAI gpt-5.6-sol-pro carries Entity-Attribute-Value triplet Model-Status-Active; xAI grok-4.5 carries Entity-Attribute-Value triplet Model-Status-Active; Anthropic claude-sonnet-5 carries Entity-Attribute-Value triplet Model-Status-Active; Kimi K2.7 carries Entity-Attribute-Value triplet Model-Status-Active; Claude Fable 5 carries Entity-Attribute-Value triplet Model-Status-Active; Qwen qwen3.7-plus carries Entity-Attribute-Value triplet Model-Status-Active; MiniMax M3 carries Entity-Attribute-Value triplet Model-Status-Active; Claude Opus 4.8 carries Entity-Attribute-Value triplet Model-Status-Active; Qwen3.7 Max carries Entity-Attribute-Value triplet Model-Status-Active; Grok Build CLI carries Entity-Attribute-Value triplet Model-UseCase-CLI coding; Gemini 3.5 Flash carries Entity-Attribute-Value triplet Model-Status-Active; Grok 4.3 carries Entity-Attribute-Value triplet Model-Status-Active; Mistral Medium 3.5 carries Entity-Attribute-Value triplet Model-Status-Active; GPT-5.5 Pro carries Entity-Attribute-Value triplet Model-Status-Active; GPT-5.5 carries Entity-Attribute-Value triplet Model-Status-Active; DeepSeek V4 Pro carries Entity-Attribute-Value triplet Model-Status-Active; Grok 4.20 carries Entity-Attribute-Value triplet Model-Status-Active; GPT-5.3 Codex carries Entity-Attribute-Value triplet Model-Status-Active; Gemini 3.1 Pro carries Entity-Attribute-Value triplet Model-Status-Active; Claude Sonnet 4.6 carries Entity-Attribute-Value triplet Model-Status-Active; OpenAI gpt-5.6-luna carries Entity-Attribute-Value triplet Model-Status-Active; OpenAI gpt-5.6-terra carries Entity-Attribute-Value triplet Model-Status-Active; OpenAI gpt-5.6-sol carries Entity-Attribute-Value triplet Model-Status-Active; Cursor 2 carries Entity-Attribute-Value triplet Model-UseCase-Coding IDE; Claude Code carries Entity-Attribute-Value triplet Model-UseCase-Coding assistant; Aider carries Entity-Attribute-Value triplet Model-UseCase-CLI coding tool; Gemini CLI carries Entity-Attribute-Value triplet Model-UseCase-CLI coding; Windsurf carries Entity-Attribute-Value triplet Model-UseCase-Coding IDE; Cline carries Entity-Attribute-Value triplet Model-UseCase-CLI coding; OpenAI Codex CLI carries Entity-Attribute-Value triplet Model-UseCase-CLI coding.
Grok 3 carries retired status while Grok 4.5, Grok 4.3, Grok 4.20, and Grok Build CLI hold active frontier classification. All current xAI models list unverified pricing. Grok Build CLI alone carries the explicit coding-CLI use-case attribute.
| Model | Status | Primary Attribute | Pricing | Use Case |
|---|---|---|---|---|
| Grok 3 | Retired | General chat (historical) | N/A | Historical only |
| Grok 4.5 | Active | Latest general-purpose LLM | Unverified | Broad queries |
| Grok 4.3 | Active | Frontier LLM | Unverified | Broad queries |
| Grok 4.20 | Active | Frontier LLM | Unverified | Broad queries |
| Grok Build CLI | Active | CLI coding interface | Unverified | Developer workflows |
OpenAI gpt-5.6-luna-pro, OpenAI gpt-5.6-terra-pro, and Anthropic claude-sonnet-5 occupy the same active row with identical unverified pricing fields. Cursor 2 and Aider supply separate coding-tool rows. No model in the table supplies documented token limits or latency values.
| Model | Status | Primary Attribute | Pricing | Use Case |
|---|---|---|---|---|
| DeepSeek deepseek-v4-flash-0731 | Active | Frontier LLM | Unverified | Broad queries |
| Qwen qwen3.7-flash | Active | Frontier LLM | Unverified | Broad queries |
| Anthropic claude-opus-5 | Active | Frontier LLM | Unverified | Broad queries |
| Cursor 2 | Active | Coding IDE | Unverified | Developer workflows |
| Claude Code | Active | Coding assistant | Unverified | Developer workflows |
| Aider | Active | CLI coding tool | Unverified | Developer workflows |
| OpenAI gpt-5.6-luna-pro | Active | Frontier LLM | Unverified | Broad queries |
| Gemini 3.1 Pro | Active | Frontier LLM | Unverified | Broad queries |
| GPT-5.5 Pro | Active | Frontier LLM | Unverified | Broad queries |
| Moonshot kimi-k3 | Active | Frontier LLM | Unverified | Broad queries |
| OpenAI gpt-5.6-terra-pro | Active | Frontier LLM | Unverified | Broad queries |
| OpenAI gpt-5.6-sol-pro | Active | Frontier LLM | Unverified | Broad queries |
| Kimi K2.7 | Active | Frontier LLM | Unverified | Broad queries |
| Claude Fable 5 | Active | Frontier LLM | Unverified | Broad queries |
| Qwen qwen3.7-plus | Active | Frontier LLM | Unverified | Broad queries |
| MiniMax M3 | Active | Frontier LLM | Unverified | Broad queries |
| Claude Opus 4.8 | Active | Frontier LLM | Unverified | Broad queries |
| Qwen3.7 Max | Active | Frontier LLM | Unverified | Broad queries |
| Gemini 3.5 Flash | Active | Frontier LLM | Unverified | Broad queries |
| Mistral Medium 3.5 | Active | Frontier LLM | Unverified | Broad queries |
| GPT-5.5 | Active | Frontier LLM | Unverified | Broad queries |
| DeepSeek V4 Pro | Active | Frontier LLM | Unverified | Broad queries |
| GPT-5.3 Codex | Active | Frontier LLM | Unverified | Broad queries |
| Claude Sonnet 4.6 | Active | Frontier LLM | Unverified | Broad queries |
Researchers must select only models listed as active on the 2026-08-01 frontier roster. Retired entries such as Grok 3 serve historical analysis exclusively. All frontier LLMs carry unverified pricing, requiring direct vendor confirmation before deployment.
Prioritize verified active tools only. The roster contains 28 entries including DeepSeek V4 Pro, Qwen3.7 Max, MiniMax M3, Mistral Medium 3.5, Gemini 3.5 Flash, and GPT-5.5 Pro. Each entry receives the same “unverified” pricing tag. Researchers cross-reference the list before any benchmark citation.
Use retired models like Grok 3 solely for historical analysis. The Grok 3 Review 2026: Why Researchers Should Upgrade from the Retired Model article supplies additional archival context. Compare against current active frontier tools with unverified details noted explicitly.
Actionable comparison steps include:
Extract the exact version string from the 2026-08-01 list.
Record the designated use-case attribute (general LLM or CLI coding tool).
Flag every missing numeric attribute such as price or context length.
Link to active documentation for Grok 4.5 or Grok Build CLI.
These steps maintain accuracy across 28 active models and exclude all retired entries. The 28-model roster breaks into 24 general-purpose LLMs and 4 coding-specific tools. Every general-purpose entry carries Entity-Attribute-Value triplet Pricing-Unverified. Every coding tool carries Entity-Attribute-Value triplet UseCase-Developer workflows.
No, Grok 3 is explicitly listed as a retired model in the 2026-08-01 landscape and should only be referenced historically.
Active frontier options include Grok 4.5, Grok 4.3, Grok 4.20, and the coding-focused Grok Build CLI.
No benchmarks or statistics are documented for Grok 3 or any models in the provided landscape.
Treat them as historical context only and compare against current active frontier tools with unverified details noted.
It serves as the CLI-focused coding tool among xAI's active offerings for developers.
Explore more AI tools and guides
I spend $5,000+ monthly on AI subscriptions so you don’t have to. Every review comes from hands-on experience — not marketing claims.
Grok 3 is retired and obsolete in 2026. This researcher-focused review explains the shift to the active Grok 4.x lineup and delivers direct comparisons against Claude Opus 5, GPT-5.6-luna-pro and other frontier tools for serious AI work.
Explore how top frontier models power AI chatbot for customer service solutions in 2026. This review delivers researcher-centric analysis of capabilities, gaps, and testing approaches for real deployments.
Discover which frontier LLMs deliver the best free chatbot for website experiences in 2026. We benchmark integration ease, latency, and real-world free-tier constraints for business deployments.
Join builders getting hands-on AI tool analysis — never sponsored, always tested.