Anthropic makes Claude, OpenAI makes ChatGPT. Both now ship a full model family — a top model, a fast mid-tier model and a cheap budget model — and both offer a free chat app plus a $20/month plan. Here is how the two line-ups compare on benchmarks, API prices, free access and features.
Anthropic leads on overall intelligence and coding: Claude Opus 5.5 is #1 on the Artificial Analysis Intelligence Index (58) and Claude Sonnet 5.5 is #2 (56), ahead of GPT-6 Astra (53). OpenAI leads on hard math — GPT-6 Astra scores 97.6% on FrontierMath Tier 4 — and ChatGPT can also create images with GPT Image 2.5. On price the two are level in the mid-tier and budget classes, while Anthropic’s top model is the cheaper flagship API.
| Category | Winner | Why |
|---|---|---|
| Overall intelligence | Anthropic | Claude Opus 5.5 is #1 on the Artificial Analysis Intelligence Index (58) and Claude Sonnet 5.5 is #2 (56); GPT-6 Astra scores 53. |
| Coding | Anthropic | Claude Opus 5.5 reaches 89.9% on SWE-bench Pro, and Claude Sonnet 5.5 scores 70.6% on Terminal-Bench 4.0 versus 57.9% for GPT-6 Astra. |
| Math & science | OpenAI | GPT-6 Astra scores 97.6% on FrontierMath Tier 4, the hardest public math benchmark. |
| Free chat app | Anthropic | Claude’s free plan runs Claude Sonnet 5.5 (#2 on the AA Intelligence Index); ChatGPT Free defaults to the budget model GPT-6 Luna. Usage limits apply to both. |
| Image generation | OpenAI | ChatGPT creates images with GPT Image 2.5, #1 on the Artificial Analysis text-to-image arena. Claude does not generate images. |
| Flagship API price | Anthropic | Claude Opus 5.5 costs $4/$20 per 1M input/output tokens; GPT-6 Astra costs $10/$50. |
| Mid-tier API price | Tie | Claude Sonnet 5.5 and GPT-6.1 Sol both cost $2/$10 per 1M tokens. |
| Cheapest API | Tie | Claude Haiku 5.5 and GPT-6 Luna both cost $0.10/$0.50 per 1M tokens. |
| Model | Score | AA Index | API price (per 1M in/out) | Context | Free in app |
|---|---|---|---|---|---|
| Claude Opus 5.5 | 99 | 58 | $4/$20 | 1M | No |
| Claude Fable 5.1 | 98 | 53 | $10/$50 | 1M | No |
| Claude Sonnet 5.5 | 97 | 56 | $2/$10 | 1M | Yes |
| Claude Haiku 5.5 | 91 | 43 | $0.10/$0.50 | 1M | Yes |
| Model | Score | AA Index | API price (per 1M in/out) | Context | Free in app |
|---|---|---|---|---|---|
| GPT-6 Astra | 98 | 53 | $10/$50 | 1M | No |
| GPT-6.1 Sol | 96 | 52 | $2/$10 | 1.05M | No |
| GPT-6 Luna | 87 | 38 | $0.10/$0.50 | 1M | Yes |
| GPT-4o | 80 | — | $2.50/$10 | 128K | No |
Score = AIVerso score (0–100). AA Index = Artificial Analysis Intelligence Index where published. API prices are list prices in USD per 1M input/output tokens.
🔗 Affiliate link — we may earn a commission if you sign up, at no extra cost to you.
| Feature | Claude | ChatGPT |
|---|---|---|
| Web search | ✓ | ✓ |
| Voice mode | ✓ | ✓ |
| PDF & file analysis | ✓ | ✓ |
| Image generation | ✗ | ✓ |
| iOS & Android apps | ✓ | ✓ |
It depends on the job. Anthropic’s Claude models lead the Artificial Analysis Intelligence Index (Claude Opus 5.5: 58, Claude Sonnet 5.5: 56, GPT-6 Astra: 53) and are the stronger choice for coding and long, careful writing. OpenAI’s GPT-6 Astra leads on hard math (97.6% on FrontierMath Tier 4), and ChatGPT adds image generation with GPT Image 2.5.
On current benchmarks, yes. Claude Opus 5.5 reaches 89.9% on SWE-bench Pro, and on Terminal-Bench 4.0 Claude Sonnet 5.5 scores 70.6% and Claude Opus 5.5 66.4%, versus 57.9% for GPT-6 Astra. For most real projects, the editor or agent you use matters as much as the model.
The apps cost the same: Claude Pro and ChatGPT Plus are both $20/month, and both have a free plan. Via the API, the mid-tier models (Claude Sonnet 5.5 and GPT-6.1 Sol) both cost $2/$10 per 1M tokens and the budget models (Claude Haiku 5.5 and GPT-6 Luna) both cost $0.10/$0.50; the flagship Claude Opus 5.5 ($4/$20) is cheaper than GPT-6 Astra ($10/$50).
Yes. Claude’s free plan gives you Claude Sonnet 5.5 on claude.ai, and ChatGPT Free uses GPT-6 Luna by default. Both free plans have usage limits.
Both are US AI companies. Anthropic focuses on its Claude models for chat, writing, coding and agents. OpenAI makes ChatGPT and the GPT models, and also offers image generation (GPT Image 2.5) and speech recognition (Whisper).
Both line-ups support about 1 million tokens: Claude Opus 5.5, Sonnet 5.5, Haiku 5.5, GPT-6 Astra and GPT-6 Luna have a 1M-token context window, and GPT-6.1 Sol has 1.05M.
How we compare: Benchmark results come from Artificial Analysis and the providers’ published results; prices and plans from the providers’ pricing pages. Commissions never influence our verdicts. Last updated: October 23, 2026.