Comparison
Claude vs GPT-5: which frontier model should you pay for?
Anthropic's flagship Claude against OpenAI's GPT-5 line on price, context and declared capabilities, with live figures.
Editorial updated 21 Sept 2026
Source: OpenRouter (live)
This is the comparison most teams actually make. Both labs ship a top-tier model that is competent at nearly everything, so the decision usually comes down to price at your traffic mix, context needs, and which one behaves better in your specific workflow.
The table below is live: prices, context window and capabilities are read from the current OpenRouter listing each time the page renders, so the figures will not be stale. Capabilities are what each provider declares on the listing; “Not declared” means the feature is not advertised there, not that it is absent.
In our own testing Claude is the steadier choice for long agentic sessions and careful writing, while GPT-5 tends to win on precise, surgical edits and structured output. Neither advantage is large enough to override a 2–3× price difference if your workload is high-volume.
Live figures
The numbers
Best value in each row is highlighted. “Not declared” means the provider does not advertise the feature on its listing.
| Attribute | ||
|---|---|---|
| Input priceUSD per 1M tokens | $5.00 | $0.20 (best) |
| Output priceUSD per 1M tokens | $25.00 | $1.20 (best) |
| Blended price75% in / 25% out | $10.00 | $0.45 (best) |
| Vs. market median | 12.5× dearer | 44% cheaper |
| Value rankcheapest = 1 | #384 | #143 (best) |
| Context windowtokens | 1M | 1.1M (best) |
| Max outputtokens | 128K | 128K |
| Input modalities | text, image, file | file, image, text |
| Output modalities | text | text |
| Tool calling | Declared | Declared |
| Reasoning | Declared | Declared |
| Structured output | Declared | Declared |
| Listed on OpenRouter | 2026-07-24 (best) | 2026-07-09 |
Blended price assumes a 3:1 input-to-output token mix. Median is across all 417 priced models.
Add or swap modelsRead-out
What the numbers say
- OpenAI: GPT-5.6 Luna Pro is the cheapest here at $0.45/M blended — 22× less than Anthropic: Claude Opus 5.
- OpenAI: GPT-5.6 Luna Pro offers the largest context window (1.1M tokens) versus 1M for Anthropic: Claude Opus 5.
Our view
Which should you choose?
- Claude Opus 5
- Choose Claude if you run long multi-step coding or research sessions and value consistency over raw speed.
- GPT-5.6 Luna Pro
- Choose GPT-5 if you need the broadest tool ecosystem, tight JSON output, or the lower per-token price at your traffic mix.
Models
Model pages
Full pricing history, capabilities and alternatives for each model in this comparison.
Anthropic
$10.00
/M blended
OpenAI
$0.450
/M blended
More comparisons
Other head-to-heads
- Claude Sonnet vs GPT-5 Mini: the mid-tier workhorses
The two models most production apps actually run. How the price gap compares to the capability gap.
- Gemini vs GPT-5: Google's flagship against OpenAI's
Gemini's very long context and multimodal input against GPT-5's ecosystem and pricing.
- Claude vs Gemini: Anthropic's flagship against Google's
Two very different philosophies: Claude's careful, consistent output versus Gemini's scale and context.
Newsletter
Stay Ahead of AI
Get the most important developments in artificial intelligence delivered directly to your inbox.
No hype. No spam.
Just trusted insights, major model releases, pricing updates, benchmark changes, product reviews, and practical guidance from across the AI ecosystem.
- Weekly AI Briefing
- Major Model Releases
- Pricing & Benchmark Updates
- Unsubscribe Anytime
The AI HQ Briefing
One email a week. Read in five minutes.