Claude Opus 5 vs Grok 4.7

Where it leads

• Higher Arena Text score, 1492 Elo vs 1441 Elo
• Higher Arena WebDev score, 1691 Elo vs 1637 Elo
• Higher FrontierMath Tiers 1-3 (v2) score, 85.6% vs 53%
• Higher GPQA Diamond score, 93.9% vs 92.7%
• Higher SimpleQA Verified score, 59.9% vs 56%
• Larger context window 1,000,000 vs 500,000 tokens
• Accepts image input
• Batch pricing, $2.50 input and $12.50 output per 1M tokens

Best for

• Complex agentic coding and enterprise work as the general-purpose default before reaching for Fable 5.1
• Latency-sensitive work that can justify Fast mode pricing

Watch out for

• Fast mode pricing applies to the whole request, including above 200K input tokens, so it doubles the effective rate rather than adding a surcharge
• Cache reads are $0.50 per million (the standard 0.1x), five times Fable 5.1’s $0.25, even though Opus 5 is the cheaper model on base tokens. A cache-heavy agent loop can invert the expected cost ordering between the two.

Where it leads

• Lower input and output prices, $2 and $6 vs $5 and $25 per 1M tokens

Best for

• Output-heavy generation under the 200,000-token prompt line
• A cheaper alternative to the frontier tiers where 500k of context is enough

Watch out for

• The long-prompt threshold is 200,000 tokens, the lowest of any model here, and reaching it doubles input and output for the entire request
• xAI's model page does not state a maximum output for Grok 4.7, so it is recorded as unknown rather than inferred
• That page does not explicitly list image input for this model either, so it is catalogued as text-only