Agnes 2.5 Pro Theta vs Step 5 Preview

Which is better for the money: Agnes 2.5 Pro Theta by Sapiens AI or Step 5 Preview by StepFun? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-11-09.

Trade-off
Sapiens AIAgnes 2.5 Pro ThetaOff the frontier
Intelligence
24.00
Total cost
$397
StepFunStep 5 PreviewOff the frontier
Intelligence
28.50
Total cost
$1644
Verdict

Depends on your budget

Step 5 Preview scores 4.5 points higher (28.5 vs 24.0) but costs 4.1× as much to run the benchmark ($1644 vs $397; premium vs mid-range cost classes). They are not alternatives at one budget: pick Step 5 Preview when you need intelligence above 24.0, and Agnes 2.5 Pro Theta when 24.0 is enough.

MetricAgnes 2.5 Pro ThetaStep 5 PreviewDifference
Intelligence (higher is better)24.028.5+18.8% Step 5 Preview is 18.8% more intelligent
Total AA cost (lower is better)$396.84$1,643.75+314.2% Step 5 Preview is 4.1× more expensive
End-to-end time (lower is better)12.07s17.62s+46.0% Step 5 Preview is 46.0% slower
Price per 1M tokens (in / out)$0.10 / $0.30$1.00 / $2.70
Input modalitiesText, ImageText, Image

Step 5 Preview vs Agnes 2.5 Pro Theta: 18.8% more intelligence, 314.2% more cost to run the AA benchmark, and 46.0% slower end-to-end.

Open this matchup in the interactive pickerAgnes 2.5 Pro Theta: what beats itStep 5 Preview: what beats it
Today's Pareto frontier

Against the frontier

0204060$10$20$50$100$200$500$1k$2k$5k$10k$20kCost to run the Artificial Analysis Intelligence Index (USD, log scale)Intelligence IndexBeam 1.0 (Reflection): intelligence 17.70, $250Claude 4.5 Haiku (Reasoning) (Anthropic): intelligence 7.40, $541Claude Fable 5.1 (High, Default Fallback) (Anthropic): intelligence 47.20, $6618Claude Fable 5.1 (Low, Default Fallback) (Anthropic): intelligence 38.70, $3886Claude Fable 5.1 (Max, Default Fallback) (Anthropic): intelligence 49.20, $13327Claude Fable 5.1 (Medium, Default Fallback) (Anthropic): intelligence 43.20, $4897Claude Fable 5.1 (Xhigh, Default Fallback) (Anthropic): intelligence 50.00, $10097Claude Haiku 5.5 (Max) (Anthropic): intelligence 36.50, $339Claude Haiku 5.5 (Xhigh) (Anthropic): intelligence 34.60, $193Claude Opus 5 (Max) (Anthropic): intelligence 44.20, $9010Claude Opus 5.5 (High, Default Fallback) (Anthropic): intelligence 51.80, $2716Claude Opus 5.5 (Low, Default Fallback) (Anthropic): intelligence 34.80, $980Claude Opus 5.5 (Max, Default Fallback) (Anthropic): intelligence 59.10, $8910Claude Opus 5.5 (Medium, Default Fallback) (Anthropic): intelligence 48.20, $2031Claude Sonnet 5.5 (High, Default Fallback) (Anthropic): intelligence 41.50, $1268Claude Sonnet 5.5 (Low, Default Fallback) (Anthropic): intelligence 25.60, $499Claude Sonnet 5.5 (Max, Default Fallback) (Anthropic): intelligence 54.50, $7981Claude Sonnet 5.5 (Medium, Default Fallback) (Anthropic): intelligence 30.80, $679Claude Sonnet 5.5 (Xhigh, Default Fallback) (Anthropic): intelligence 49.90, $2903DeepSeek V4 Pro 0813 (Max) (DeepSeek): intelligence 23.10, $1402DeepSeek V4 Pro 0813 (Non-reasoning) (DeepSeek): intelligence 12.60, $771DeepSeek V4.1 Flash 1017 (High) (DeepSeek): intelligence 28.60, $384DeepSeek V4.1 Flash 1017 (Max) (DeepSeek): intelligence 32.60, $521DeepSeek V4.1 Flash 1017 (Non-reasoning) (DeepSeek): intelligence 18.50, $310Gemini 3.1 Flash-Lite (Google): intelligence 6.20, $224Gemini 3.5 Flash-Lite (Google): intelligence 11.70, $654Gemini 3.8 Flash (High) (Google): intelligence 30.80, $5764Gemini 3.8 Flash (Medium) (Google): intelligence 30.10, $4547Gemini 4 Argon (High) (Google): intelligence 50.40, $7620Gemini 4 Carbon (Low) (Google): intelligence 54.20, $5725GLM-5.3 (Low) (Z AI): intelligence 23.10, $1219GLM-5.3 (Max) (Z AI): intelligence 31.70, $3496GLM-5.3-Flash (Z AI): intelligence 27.80, $378GPT-5.6 Luna (Max) (OpenAI): intelligence 23.80, $334GPT-5.6 Terra (Max) (OpenAI): intelligence 32.30, $3250GPT-6 Astra (Low) (OpenAI): intelligence 43.30, $1651GPT-6 Astra (Max) (OpenAI): intelligence 56.90, $5661GPT-6 Astra (Medium) (OpenAI): intelligence 51.60, $2638GPT-6 Luna (Max) (OpenAI): intelligence 27.90, $136GPT-6 Luna (Non-reasoning) (OpenAI): intelligence 9.40, $53.1GPT-6 Luna (Xhigh) (OpenAI): intelligence 23.50, $85.7GPT-6 Sol (Max) (OpenAI): intelligence 44.20, $1930GPT-6.1 Sol (Low) (OpenAI): intelligence 34.80, $238GPT-6.1 Terra (Max) (OpenAI): intelligence 51.30, $972GPT-6.1 Terra (Non-reasoning) (OpenAI): intelligence 30.50, $213GPT-6.1 Terra (Xhigh) (OpenAI): intelligence 45.30, $400Granite 4.2 8B (IBM): intelligence 5.10, $46.9Grok 4.6 (Xhigh) (SpaceXAI): intelligence 31.00, $3391Grok 4.7 (High) (SpaceXAI): intelligence 35.20, $4722Grok 4.7 (Low) (SpaceXAI): intelligence 31.10, $2169Grok 4.7 (Xhigh) (SpaceXAI): intelligence 35.50, $5906Inkling (Xhigh) (Thinking Machines): intelligence 12.50, $1237Inkling 2 Small (Thinking Machines): intelligence 30.60, $792Inkling Small (Thinking Machines): intelligence 12.90, $225Kimi K3 (Low) (Kimi): intelligence 17.40, $1764Kimi K3.1 Pro (Kimi): intelligence 46.50, $3153Ling 3.1 Flash 1019 (InclusionAI): intelligence 29.60, $2009MiMo-V2.6-Pro (Xiaomi): intelligence 29.90, $221MiniMax-M3 (MiniMax): intelligence 18.90, $878MiniMax-M3.1-Flash (MiniMax): intelligence 29.10, $358Mistral Large 4 (High) (Mistral): intelligence 26.10, $1360Mistral Large 4 (Medium) (Mistral): intelligence 23.10, $936Mistral Large 4 Preview (Mistral): intelligence 25.30, $2923Mistral Medium 3.5 (Mistral): intelligence 4.60, $1103Mistral Small 4 (Reasoning) (Mistral): intelligence 4.30, $43.4Muse Glimmer (High) (Meta): intelligence 6.30, $182Muse Spark 1.3 (Max) (Meta): intelligence 36.60, $1906Muse Spark Open (Xhigh) (Meta): intelligence 24.90, $1697Muse Wonder (High) (Meta): intelligence 40.50, $3520Muse Wonder (Max) (Meta): intelligence 43.60, $5414Nemotron 3 Ultra 550B A55B (Reasoning) (NVIDIA): intelligence 11.80, $1542Nemotron 3.5 Lightning (NVIDIA): intelligence 5.50, $227Nemotron 3.5 Super VL (NVIDIA): intelligence 13.60, $447Qwen3.8 27B (Low) (Alibaba): intelligence 15.70, $1460Qwen3.8 27B (Medium) (Alibaba): intelligence 17.50, $1636Qwen3.8 27B (Non-reasoning) (Alibaba): intelligence 10.00, $2642Qwen3.8 27B (Xhigh) (Alibaba): intelligence 20.10, $2390Qwen3.8 Max (0902) (Alibaba): intelligence 33.70, $6354Tencent Hy4 (Tencent): intelligence 34.60, $1021GPT-6 Luna (Low) (OpenAI): intelligence 9.70, $9.09GPT-6 Luna (Medium) (OpenAI): intelligence 18.50, $39.5Claude Haiku 5.5 (Low) (Anthropic): intelligence 18.60, $39.6GPT-6 Luna (High) (OpenAI): intelligence 21.40, $62.6Claude Haiku 5.5 (Medium) (Anthropic): intelligence 24.60, $72.7MiMo-V2.6-Flash (Xiaomi): intelligence 26.70, $106GPT-6.1 Terra (Low) (OpenAI): intelligence 29.40, $111Claude Haiku 5.5 (High) (Anthropic): intelligence 29.90, $121GPT-6.1 Terra (Medium) (OpenAI): intelligence 38.20, $190GPT-6.1 Terra (High) (OpenAI): intelligence 42.40, $267GPT-6.1 Sol (Medium) (OpenAI): intelligence 46.30, $389GPT-6.1 Sol (High) (OpenAI): intelligence 50.30, $573GPT-6.1 Sol (Xhigh) (OpenAI): intelligence 52.50, $718GPT-6.1 Sol (Max) (OpenAI): intelligence 54.60, $1124GPT-6 Astra (High) (OpenAI): intelligence 54.80, $3204GPT-6 Astra (Xhigh) (OpenAI): intelligence 55.30, $3746Claude Opus 5.5 (Xhigh, Default Fallback) (Anthropic): intelligence 57.40, $4766Gemini 4 Carbon (Medium) (Google): intelligence 58.30, $5903Gemini 4 Carbon (High) (Google): intelligence 60.00, $8240Step 5 Preview (StepFun): intelligence 28.50, $1644Agnes 2.5 Pro Theta (Sapiens AI): intelligence 24.00, $397Agnes 2.5 Pro ThetaStep 5 PreviewGPT-6 Luna (Low)Gemini 4 Carbon (High)
Agnes 2.5 Pro ThetaStep 5 PreviewToday's Pareto frontierOpen-weights frontierEvery other model
Matchup

Compare something else

Agents: Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/agnes-2-5-pro-theta/step-5 · Markdown: /api/v1/compare/agnes-2-5-pro-theta/step-5?format=md · llms.txt · agent skill