Grok 3
Feb 2025
Retired 2026-05-15Replace withGrok 4.7Newer version
Gemini 4 ArgonCheaper, as good
GPT-6 LunaBetter
MiMo v2.6 ProOpen weights
Grok 4.7Newer versionGemini 4 ArgonCheaper, as goodGPT-6 LunaBetterMiMo v2.6 ProOpen weightsGrok 3Compare inon
Benchmarks of Grok 320
BenchmarkByScoreRank
Overall1
Knowledge5
Reasoning4
Math1
Science2
Coding1
Health & law1
Documents1
Legacy · saturated or no longer reported4
No benchmark matches
Rankingin non-thinking mode?
Retired models are left out of the current rankings.
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperClaude Sonnet 5.5max×4.6-39%$5.46No forced tool use+5BetterGPT-6 Lunalow+78%-97%$0.005Function calling only with none · Chat Completions+3OpenMiMo v2.6 Prothinking×3.8-92%$0.13Pass back reasoning_content+3
CheaperKimi K3.1 Prounstated×4.5-39%$2.11CheaperGemini 4 Argonhigh×4.3-39%*$1.62Synchronous output stops at 65,536+2ModelScoreCostPer taskWhat changes
Claude Sonnet 5.5max×4.6-39%$5.46No forced tool use+5Kimi K3.1 Prounstated×4.5-39%$2.11Gemini 4 Argonhigh×4.3-39%*$1.62Synchronous output stops at 65,536+2GPT-6.1 Solmax×4.3-39%$0.72Remove unsupported parameters+5GPT-6.1 Terramax×4-70%$0.40ModelScoreCostPer taskWhat changes
GPT-6 Lunalow+78%-97%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow+74%-93%$0.01Claude Haiku 5.5low×2.4-97%$0.024Remove sampling parameters+5
Ling 3.0 FlashthinkingOpen+66%-99%$0.028MiMo v2.5thinkingOpen+80%-97%$0.042ModelScoreCostPer taskWhat changes
MiMo v2.6 ProthinkingOpen×3.8-92%$0.13Pass back reasoning_content+3
DeepSeek V4.1 Flash 1017maxOpen×3.7-92%$0.26Pass back reasoning_content · Chat Completions+4
GLM 5.3maxOpen×3.7-67%$2.01Reasoning cannot be disabled+3
Step 5 PreviewthinkingOpen×3.6-78%$1.03web_search not available · Chat Completions+2Kimi K3maxOpen×3.6-9%$2.00Sampling parameters are fixed+3ModelScoreCostPer taskWhat changes
Grok 4.6medium×3.5—$1.14Grok 4.7low×3.5—$1.03Reasoning cannot be disabled+4Grok 4.5high×3.2—$0.76Grok Build 0.1thinking×2.2—$1 / $2Grok 4.3medium×2—$1.25 / $2.50Grok 4.20 Multi-Agentn/a—$1.25 / $2.50Grok 4.20 Non-Reasoningn/a—$1.25 / $2.50Context: not verifiedGrok 4.20 Reasoningn/a—$1.25 / $2.50Context: not verifiedAll xAI models →History
8 versionsBenchmarks
all categoriesChat & Knowledge
AA Omniscience Index-33.2
GPQA Diamond69.3
Coding
Agents & Tool Use
τ²-bench · AA48.8
Writing & Language
FAQ
How much does Grok 3 cost?›
$3.30 per 1M input tokens and $16.50 per 1M output tokens (cheapest host: Azure AI Foundry).
Is Grok 3 free to use?›
There is no free API; the cheapest paid price is $3.30 per 1M input tokens.
How can I use Grok 3?›
Through the API of Azure AI Foundry.
What is the context window of Grok 3?›
131K tokens.
When was Grok 3 released? Is there a newer version?›
Released 2025-02-19. It has been superseded by Grok 4.3.
Can Grok 3 read images?›
No, it accepts text only.
What is Grok 3?›
Grok 3 is a language model from xAI, released Feb 2025.
What are the best alternatives to Grok 3?›
Claude Sonnet 5.5 (cheaper), GPT-6 Luna (better), MiMo v2.6 Pro (open weights).