TokenRate

Claude Opus 5 vs Gemini 3.1 Pro

Anthropic's July 2026 flagship against Google's GA flagship, which costs less than half as much. Compare SWE-bench, ARC-AGI-2, context pricing tiers, and which is the better value.

Comparing 2 models · Prices last verified

Verdict

Opus 5 wins on capability, Gemini 3.1 Pro wins on price — and the gap is wide on both counts. Opus 5 leads SWE-bench Verified (96.0% vs 80.6%), SWE-bench Pro (79.2% vs ~54.2%) and ARC-AGI-2 (90.4% at max effort vs 77.1%), which is unsurprising given Gemini 3.1 Pro shipped five months earlier in February 2026. But Gemini costs $2/$12 per million tokens against Opus 5's $5/$25 — less than half — and matches its 1M context with native video and audio input Opus 5 lacks. Watch Google's tiered pricing though: above 200K context it doubles to $4/$18, which erases much of the advantage on long-context work. Take Opus 5 when the hardest requests decide the outcome; take Gemini 3.1 Pro for high-volume multimodal work under 200K tokens.

Pricing Comparison

ModelProviderInput / 1MOutput / 1MContextTier
Claude Opus 5Anthropic$5.00$25.001Mflagship
Gemini 3.1 ProGoogle$2.00$12.001Mflagship

Model Breakdown

Claude Opus 5

Anthropic

$5.00

per 1M input

Claude Opus 5 is Anthropic's flagship model, released July 24, 2026. It reaches close to Claude Fable 5's frontier intelligence at half the price, and holds the same $5/$25 per-million pricing as Opus 4.8 while more than doubling its agentic coding score. It tops Frontier-Bench v0.1 (43.3%), ARC-AGI-3 (30.2%, roughly 4x the next best model), and Artificial Analysis' GDPval-AA v2 knowledge-work leaderboard (1,861 Elo). It is also Anthropic's most aligned Opus model to date.

Near-Fable-5 frontier quality at half the price ($5/$25 vs $10/$50)
State of the art on agentic coding, computer use, and knowledge work

$2.00

per 1M input

Gemini 3.1 Pro is Google's generally-available flagship (released February 19, 2026) for complex reasoning, long-context work, and native multimodality across text, images, video, audio, and PDFs. It posts 80.6% on SWE-bench Verified, 94.3% on GPQA Diamond, and 77.1% on ARC-AGI-2, with a 1M-token context window at $2/$12 per million tokens — materially cheaper than the Claude and GPT flagships. Despite persistent speculation, there is no Gemini 3.5 Pro: Google shipped only Flash-class models in July 2026, leaving 3.1 Pro as its top Pro tier.

Frontier reasoning at well under flagship pricing ($2/$12 per 1M)
1M-token context with native video, audio, image and PDF input

Our Verdict

Opus 5 wins on capability, Gemini 3.1 Pro wins on price — and the gap is wide on both counts. Opus 5 leads SWE-bench Verified (96.0% vs 80.6%), SWE-bench Pro (79.2% vs ~54.2%) and ARC-AGI-2 (90.4% at max effort vs 77.1%), which is unsurprising given Gemini 3.1 Pro shipped five months earlier in February 2026. But Gemini costs $2/$12 per million tokens against Opus 5's $5/$25 — less than half — and matches its 1M context with native video and audio input Opus 5 lacks. Watch Google's tiered pricing though: above 200K context it doubles to $4/$18, which erases much of the advantage on long-context work. Take Opus 5 when the hardest requests decide the outcome; take Gemini 3.1 Pro for high-volume multimodal work under 200K tokens.

FAQ

Compare These Models Yourself

Use the TokenRate calculator to enter your budget or token count and see the exact cost for each model side by side.

Open Calculator →

More Comparisons