Leaderboards
Benchmarks
Every score carries its origin.
Arena Agent55Arena Agent · Recovery55Arena Agent · Steerability55Arena Agent · Task outcome55Arena Agent · Tool use55LiveBench Agentic Coding58SWE-bench Verified42
LiveBench Agentic Coding
% average over LiveBench agentic coding tasks, run inside an agent harnessProvenance58 modelsscores Jun 25, 2026
1DeepSeek V4.1 FlashDeepSeek763BScore77.27Score / $193Score / GB VRAM~0.16ScoredJun 25, 20262Claude Opus 5.5AnthropicScore71.72Score / $4ScoredJun 25, 20263Claude Fable 5.1AnthropicScore66.06Score / $1ScoredJun 25, 20264Claude Opus 5AnthropicScore65.2Score / $3ScoredJun 25, 20265DeepSeek V4 Flash Vision ExpDeepSeek305BScore65.1Score / $101Score / GB VRAM~0.34ScoredJun 25, 20266Qwen3 MaxQwenScore64.65Score / $17ScoredJun 25, 20267Qwen3.8 MaxQwenScore64.65Score / $13ScoredJun 25, 20268Muse Spark 1.3MetaScore64.09Score / $15ScoredJun 25, 20269Claude Fable 5AnthropicScore62.17Score / $1ScoredJun 25, 202610Kimi K3Moonshot AI2.8TScore62.17Score / $6Score / GB VRAM~0.035ScoredJun 25, 202611Qwen3.8 FlashQwenScore61.62Score / $161ScoredJun 25, 202612Qwen3.8 27BQwen27.8BScore61.36Score / $41Score / GB VRAM~3.5ScoredJun 25, 202613GLM 5.3Z.aiScore60.91Score / $53ScoredJun 25, 202614Claude Sonnet 5AnthropicScore59.39Score / $6ScoredJun 25, 202615Muse Spark 1.1MetaScore58.54Score / $14ScoredJun 25, 202616Gemini 3.7 FlashGoogleScore58.28Score / $31ScoredJun 25, 202617Muse Spark 1.2MetaScore57.58Score / $14ScoredJun 25, 202618GPT-6 AstraOpenAIScore57.32Score / $2ScoredJun 25, 202619Grok 4.6xAIScore57.02Score / $10ScoredJun 25, 202620GLM 5.3 FlashZ.ai321BScore56.77Score / $227Score / GB VRAM~0.28ScoredJun 25, 202621Grok 4.5xAIScore56.46Score / $9ScoredJun 25, 202622Claude Sonnet 5.5AnthropicScore56.31Score / $6ScoredJun 25, 202623GPT-5.6 SolOpenAIScore56.21Score / $11ScoredJun 25, 202624GPT-5.6 TerraOpenAIScore54.95Score / $9ScoredJun 25, 202625Union AlphastealthScore54.7ScoredJun 25, 202626Gemini 3.8 FlashGoogleScore54.24Score / $29ScoredJun 25, 202627GPT-5.5OpenAIScore53.99Score / $4ScoredJun 25, 202628Grok 4.7xAIScore53.99Score / $9ScoredJun 25, 202629GPT-5.4OpenAIScore53.84Score / $7ScoredJun 25, 202630GPT-6 SolOpenAIScore52.88Score / $11ScoredJun 25, 202631Ox AlphastealthScore52.63ScoredJun 25, 202632GLM 5.2Z.ai753BScore51.77Score / $57Score / GB VRAM~0.11ScoredJun 25, 202633GPT-6 LunaOpenAIScore51.21Score / $205ScoredJun 25, 202634Claude Opus 4.7AnthropicScore50.66Score / $2ScoredJun 25, 202635Claude Opus 4.8AnthropicScore50.51Score / $2ScoredJun 25, 202636GPT-5.2 ChatOpenAIScore50.25Score / $7ScoredJun 25, 202637GPT-5.2-CodexOpenAIScore49.39Score / $4ScoredJun 25, 202638InklingThinking Machines952BScore49.39Score / $12Score / GB VRAM~0.082ScoredJun 25, 202639Gemini 3.5 FlashGoogleScore48.99Score / $11ScoredJun 25, 202640Claude Opus 4.6AnthropicScore48.99Score / $2ScoredJun 25, 202641GPT-5.6 LunaOpenAIScore48.43Score / $81ScoredJun 25, 202642Kimi K2.6Moonshot AI1.1TScore46.92Score / $26Score / GB VRAM~0.07ScoredJun 25, 202643GPT-5.4 NanoOpenAIScore46.77Score / $75ScoredJun 25, 202644Grok Build 0.1xAIScore45.81Score / $23ScoredJun 25, 202645Kimi K2.7 CodeMoonshot AI1.1TScore45.66Score / $15Score / GB VRAM~0.068ScoredJun 25, 202646Gemini 3.5 Flash LiteGoogleScore45.25Score / $36ScoredJun 25, 202647Qwen3.7 MaxQwenScore43.59Score / $12ScoredJun 25, 202648Gemini 3.6 FlashGoogleScore43.43Score / $23ScoredJun 25, 202649DeepSeek V4 ProDeepSeek1.6TScore42.63Score / $27Score / GB VRAM~0.042ScoredJun 25, 202650Claude Sonnet 4.6AnthropicScore42.63Score / $3ScoredJun 25, 202651GPT-5.4 MiniOpenAIScore41.67Score / $19ScoredJun 25, 202652Qwen3.6 PlusQwenScore41.36Score / $21ScoredJun 25, 202653MiniMax M3MiniMax427BScore40.66Score / $42Score / GB VRAM~0.15ScoredJun 25, 202654Claude Opus 4.5AnthropicScore39.7Score / $2ScoredJun 25, 202655Qwen3.6 27BQwen27.8BScore39.29Score / $20Score / GB VRAM~2.2ScoredJun 25, 202656Nemotron 3 UltraNVIDIA561BScore38.74Score / $18Score / GB VRAM~0.11ScoredJun 25, 202657DeepSeek V4 FlashDeepSeek291BScore37.63Score / $449Score / GB VRAM~0.21ScoredJun 25, 202658Grok 4.3xAIScore18.54Score / $7ScoredJun 25, 2026