AI Model Comparisons
Claude 5.5 family and GPT-6 models compared side by side — pricing, benchmarks and specs.
Last updated 2026-10-02 · Vendor data sourced from Anthropic and OpenAI
All Models at a Glance
| Spec | Claude Fable 5.1 | Claude Opus 5.5 | Claude Sonnet 5.5 | Claude Sonnet 5 | GPT-6 Astra | GPT-6 Luna | GPT-6 Sol | GPT-6.1 Sol |
|---|---|---|---|---|---|---|---|---|
| Vendor | Anthropic | Anthropic | Anthropic | Anthropic | OpenAI | OpenAI | OpenAI | OpenAI |
| Released | Sep 1, 2026 | Sep 22, 2026 | Sep 28, 2026 | Jun 30, 2026 | Sep 3, 2026 | Sep 22, 2026 | Sep 22, 2026 | Sep 29, 2026 |
| Input / output (per MTok) | $10 / $50 | $4 / $20 | $2 / $10 | $2 / $10 | $10 / $50 | $0.10 / $0.50 | $2 / $10 | $2 / $10 |
| Batch input / output | $5 / $25 | $2 / $10 | $1 / $5 | $1 / $5 | $5 / $25 | $0.05 / $0.25 | $1 / $5 | $1 / $5 |
| Cache read | $0.25 per MTok | $0.20 per MTok | $0.20 per MTok | $0.20 per MTok | $1 per MTok | $0.01 per MTok | $0.20 per MTok | $0.10 per MTok |
| Context window | 1M tokens | 1M tokens | 1M tokens | 1M tokens | 1.05M tokens | 1.05M tokens | 1.05M tokens | 1.05M tokens |
| Max output | 128K tokens | 128K tokens (300K on Batch API) | 128K tokens (300K on Batch API) | 128K tokens (300K on Batch API) | 128K tokens | 128K tokens | 128K tokens | 128K tokens |
| Latency | Slower | Moderate | Fast | Fast | Slower | Very fast | Fast | Fast |
| Thinking | Adaptive, always on, cannot be disabled | Adaptive, always on, cannot be disabled | Adaptive; can be tuned with between_tools (no up-front thinking at effort <= high) | Adaptive; can be disabled | Reasoning model, effort configurable | Reasoning model, effort configurable | Reasoning model, effort configurable | Reasoning model, effort configurable |
| Default effort | high | medium | high | high | medium | medium | medium | medium |
| Knowledge cutoff | Jun 2026 | Jun 2026 | Jun 2026 | Jan 2026 | Apr 2026 | May 18, 2026 | Apr 2026 | Apr 2026 |
Same-Vendor Comparisons
Claude Opus 5.5 vs Claude Sonnet 5.5
Opus leads 7 of 8 benchmarks; Sonnet wins Terminal-Bench, is faster and costs half as much.
Claude Opus 5.5 vs Claude Fable 5.1
Opus is 60% cheaper, faster and scores higher on all seven benchmarks.
GPT-6 Luna vs GPT-6 Sol
Luna is 20x cheaper and 1.7x faster; Sol leads all benchmarks but Luna nearly matches on DeepSWE.
GPT-6 Sol vs GPT-6.1 Sol
6.1 Sol scores higher on every benchmark and halves cache reads, but runs 13% slower and drops Chat Completions tools.
Claude Sonnet 5 vs Claude Sonnet 5.5
Sonnet 5.5 leads all 9 benchmarks and is 77% faster at the same $2/$10 pricing, but costs more per task at max effort.
Claude Opus 5.5 vs Claude Sonnet 5
Opus leads all 9 benchmarks and costs 4.3× less to evaluate at default effort despite 2× higher per-token pricing.
Cross-Vendor Comparisons (Anthropic vs OpenAI)
Claude Opus 5.5 vs GPT-6 Astra
Opus leads 7 of 9 and costs 60% less; Astra edges ahead on math and scientific reasoning.
Claude Opus 5.5 vs GPT-6 Sol
Opus medium outscores Sol max; Sol is 50% cheaper per token but hits a lower quality ceiling.
Browse Individual Models
Looking for one model's price, specs and cost calculator? See the AI models directory.
Model Comparison FAQ
Which Claude model should I use?
Sonnet 5.5 is the cheapest and fastest of the three and fits routine coding and everyday API work. Opus 5.5 leads most benchmarks and suits long-running agents and complex reasoning. Fable 5.1 is the premium tier at $10 / $50 per million tokens and is worth it only if your own evals show it pays off.
Claude vs GPT: which is cheaper?
Claude Opus 5.5 costs $4 / $20 per million input / output tokens and Sonnet 5.5 costs $2 / $10. On the OpenAI side, GPT-6.1 Sol matches Sonnet's $2 / $10 price point (superseding GPT-6 Sol) while GPT-6 Astra is $10 / $50. Per-token price is not the same as per-task cost, so benchmark on your own workload before deciding.
What is the cheapest Claude model?
Claude Sonnet 5.5 at $2 / $10 per million tokens. For even lower cost, Haiku 4.5 costs $1 / $5 but has a smaller 200K context window. Both offer 50% batch discounts.
What is the difference between GPT-6 Sol and GPT-6.1 Sol?
GPT-6.1 Sol scores higher on every benchmark, halves cache read costs ($0.10 vs $0.20 per MTok), and matches GPT-6 Astra on DeepSWE — all at the same $2/$10 pricing. Trade-offs: 13% slower, no Chat Completions tool calling, and drops the none/minimal effort levels.
How often is this data updated?
Specs and prices come from each vendor's official documentation and are re-verified whenever a model or price changes. Every comparison page shows its last-updated date.