Cheaper overall
Claude Opus 5
50% lower blended rate
Higher benchmark scores
GPT-6 Astra
leads on 6 of 11 figures
Bigger context window
GPT-6 Astra
1,050,000 vs 1,000,000 tokens
Cheaper cached input
Claude Opus 5
$0.50 vs $1.00, 50% less
Where each one wins
GPT-6 Astra
- Larger context window 1,050,000 tokens
- Higher overall intelligence (Intelligence Index) 52.7 vs 50.8
- Higher graduate-level science (GPQA Diamond) 96.1 vs 93.7
- Higher scientific coding (SciCode) 56.5 vs 56.4
- Ahead on 3 more benchmark figures
- Offers Flex and Priority pricing not listed for Claude Opus 5
Claude Opus 5
- Cheaper input tokens $5.00 vs $10.00, 50% less
- Cheaper output tokens $25.00 vs $50.00, 50% less
- Cheaper cached input $0.50 vs $1.00, 50% less
- Cheaper on all four workloads
- Higher coding ability (Coding Index) 78 vs 76.9
- Higher multi-step tool use (Agentic Index) 56.5 vs 51
- Higher expert-exam performance (Humanity's Last Exam) 54.9 vs 54.7
- Ahead on 2 more benchmark figures
What four workloads cost
One run, and the same run a thousand times.
Claude Opus 5 is cheaper on all four workloads, by much the same margin each time — about 2.5 times, from $0.0175 against $0.0350 on a chat turn to $2.08 against $8.23 on a long document.
| Workload | GPT-6 Astra | Claude Opus 5 | ×1,000 | Difference |
|---|---|---|---|---|
| Chat turn 1,000 in / 500 out | $0.0350 | $0.0175 | $35.00 / $17.50 | Claude Opus 5 is 50% cheaper |
| RAG answer 10,000 in / 800 out | $0.1400 | $0.0700 | $140.00 / $70.00 | Claude Opus 5 is 50% cheaper |
| Agent loop 32,000 in / 700 out × 20 calls | $1.70 | $0.8500 | $1,700.00 / $850.00 | Claude Opus 5 is 50% cheaper |
| Long document 400,000 in / 3,000 out | $8.23 | $2.08 | $8,225.00 / $2,075.00 | Claude Opus 5 is 75% cheaper |
Price per 1M tokens
Above 272,000 input tokens the rates change, and the new rate applies to the whole request, not just the tokens past that point: GPT-6 Astra's input goes from $10.00 to $20.00. It does not change which of the two is cheaper.
| Context band | Price | GPT-6 Astra | Claude Opus 5 | Difference |
|---|---|---|---|---|
| Prompts up to 272,000 tokens | Input | $10.00 | $5.00 | Claude Opus 5 is 50% cheaper |
| Prompts up to 272,000 tokens | Cached input | $1.00 | $0.50 | Claude Opus 5 is 50% cheaper |
| Prompts up to 272,000 tokens | Output | $50.00 | $25.00 | Claude Opus 5 is 50% cheaper |
| Prompts over 272,000 tokens | Input | $20.00 | $5.00 | Claude Opus 5 is 75% cheaper |
| Prompts over 272,000 tokens | Cached input | $2.00 | $0.50 | Claude Opus 5 is 75% cheaper |
| Prompts over 272,000 tokens | Output | $75.00 | $25.00 | Claude Opus 5 is 67% cheaper |
Blended: GPT-6 Astra $20.00, Claude Opus 5 $10.00 per 1M — one rate at 3:1 input to output, for comparing two models at a glance.
GPT-6 Astra priced by OpenAI, Claude Opus 5 by Anthropic. Standard tier, pay-as-you-go.
Other pricing tiers
Both models offer Batch, worth up to 50% off their standard rates, and the cheaper model is the same one on every tier — so the tier you buy changes the bill rather than the choice. Only GPT-6 Astra lists Flex and Priority, at up to 50% off its own standard rate; Claude Opus 5 offers none of them.
Every tier either model sells is here, so this is the whole price sheet. A saving is measured against that model's own standard rate — how we read tiers.
| Tier | GPT-6 Astra | Claude Opus 5 | Cheaper |
|---|---|---|---|
| Batch | $5.00 in / $25.00 out 50% off Standard | $2.50 in / $12.50 out 50% off Standard | Claude Opus 5 |
| Flex only on GPT-6 Astra | $5.00 in / $25.00 out 50% off Standard | Not offered | — |
| Priority only on GPT-6 Astra | $20.00 in / $100.00 out | Not offered | — |
Benchmarks
They share 11 figures: GPT-6 Astra leads on 6, Claude Opus 5 on 5. The widest gap is 8.3 points, on real-world task performance (GDPval).
Overall intelligence
Intelligence Index — Overall intelligence across reasoning, knowledge & math evals
GPT-6 Astra ahead by 1.9
Coding ability
Coding Index — Coding ability across software-engineering evals
Claude Opus 5 ahead by 1.1
Multi-step tool use
Agentic Index — Tool use & multi-step agent task performance
Claude Opus 5 ahead by 5.5
Graduate-level science
GPQA Diamond — Graduate-level questions in biology, chemistry & physics
GPT-6 Astra ahead by 2.4
Expert-exam performance
Humanity's Last Exam — Expert-level questions across many academic domains
Claude Opus 5 ahead by 0.2
Long-context reasoning
AA-LCR — Long-context reasoning across large inputs
Claude Opus 5 ahead by 1.3
Scientific coding
SciCode — Research-level scientific coding tasks
GPT-6 Astra ahead by 0.1
Physics research reasoning
CritPt — Unpublished physics research reasoning problems
GPT-6 Astra ahead by 2.6
Real-world task performance
GDPval — Economically valuable, real-world knowledge work
Claude Opus 5 ahead by 8.3
Factual accuracy
Omniscience Accuracy — Breadth of factual knowledge across domains
GPT-6 Astra ahead by 1.7
Answer reliability
Omniscience Non-Hallucination — How reliably the model avoids fabricated answers
GPT-6 Astra ahead by 8.2
Each figure is that model's best published run, with the effort level named beside it — where these scores come from.
Specifications
GPT-6 Astra holds 50,000 more input tokens in one request. Only GPT-6 Astra lists web search.
| Specification | GPT-6 Astra | Claude Opus 5 |
|---|---|---|
| Context window | 1,050,000 tokens | 1,000,000 tokens |
| Max output | 128,000 tokens | 128,000 tokens |
| Takes in | Text, image, file | Text, image, file |
| Puts out | Text | Text |
| Knowledge cutoff | 30 Apr 2026 | — |
| Released | 3 Sep 2026 | 24 Jul 2026 |
| Status | Active | Active |
| Sold by | OpenAI | Anthropic, Perplexity |
| Tool use | Yes | Yes |
| Structured outputs | Yes | Yes |
| Web search | Yes | Not listed |
| Prompt caching | Yes | Yes |
| Code execution | Yes | Yes |
| Computer use | Yes | Yes |
| Open weights | Not listed | Not listed |
| Reasoning | Yes | Yes |
FAQs
- Is GPT-6 Astra cheaper than Claude Opus 5?
- No — the other way round. A 1,000-token prompt with a 500-token reply costs $0.0350 on GPT-6 Astra and $0.0175 on Claude Opus 5, and the same model is cheaper on every workload on this page.
- How much do GPT-6 Astra and Claude Opus 5 cost per 1M tokens?
- GPT-6 Astra costs $10.00 for input and $50.00 for output. Claude Opus 5 costs $5.00 and $25.00. Standard pay-as-you-go rates, as of 3 Sep 2026.
- Which is better, GPT-6 Astra or Claude Opus 5?
- On published benchmarks, GPT-6 Astra. It leads on 6 of the 11 figures both models report, including overall intelligence (Intelligence Index), where it scores 52.7 against 50.8.
- Is Batch pricing cheaper for GPT-6 Astra or Claude Opus 5?
- Batch input costs $5.00 on GPT-6 Astra and $2.50 on Claude Opus 5, 50% below the standard rate on both.
- Does GPT-6 Astra or Claude Opus 5 have a bigger context window?
- GPT-6 Astra, at 1,050,000 tokens against 1,000,000. That is 50,000 more input tokens in a single request.