Key takeaways (October 2026)
The current flagship API models are OpenAI's GPT-6 Astra, Anthropic's Claude Fable 5.1 and Claude Opus 5.5, Google's Gemini 4 Argon and xAI's Grok 4.7. GPT-6 Astra and Claude Fable 5.1 both list at $10 per million input tokens and $50 per million output tokens. Claude Opus 5.5 lists at $4 / $20, and Gemini 4 Argon at $2 / $10 until November 30, 2026. Most of these models accept about 1M tokens of context.
- Largest context windows: GPT-6 Astra, Sol and Luna (1.05M tokens); the Claude 5.5 line, Claude Fable 5.1, Gemini 4 Argon and DeepSeek-V4-Pro (1M). Grok 4.7 lists 500K.
- Lowest list prices: GPT-6 Luna ($0.10 / $0.50) and Claude Haiku 5.5 (from $0.10 / from $0.50) per million input / output tokens.
- Price changes scheduled: Google lists Gemini 4 Argon rising to $4 / $20 on December 1, 2026, and Gemini 3.8 Flash rising to $1.50 / $7.50 on January 1, 2027.
- Newest knowledge cutoffs: the Claude 5.5 line and Claude Fable 5.1 (June 2026), GPT-6 Luna (May 18, 2026) and Grok 4.7 (May 2026).
- Open weights: Mistral Large 4 is open-weight. Parameter counts for the proprietary models are not published.
Comparison table
Prices are standard API list prices in USD per million tokens, checked October 16, 2026; batch, cache and regional rates differ.[*] “Not listed” means the value does not appear in the vendor's model or pricing documentation. Context window definitions vary; see methodology for details.
| Model | Vendor | API list price (input / output, USD per 1M tokens) | Context window | Input modalities | Output modalities | License | Knowledge cutoff |
|---|---|---|---|---|---|---|---|
| GPT-6 Astra | OpenAI | $10 / $50[1] | 1.05M | Text, Image | Text | Proprietary | Apr 30, 2026 |
| GPT-6.1 Sol | OpenAI | $2 / $10[2] | 1.05M | Text, Image | Text | Proprietary | Apr 30, 2026 |
| GPT-6 Luna | OpenAI | $0.10 / $0.50[3] | 1.05M | Text, Image | Text | Proprietary | May 18, 2026 |
| Claude Fable 5.1 | Anthropic | $10 / $50[4] | 1M | Text, Image | Text | Proprietary | Jun 2026 |
| Claude Opus 5.5 | Anthropic | $4 / $20[5] | 1M | Text, Image | Text | Proprietary | Jun 2026 |
| Claude Sonnet 5.5 | Anthropic | $2 / $10[6] | 1M | Text, Image | Text | Proprietary | Jun 2026 |
| Claude Haiku 5.5 | Anthropic | From $0.10 / from $0.50[7] | 1M | Text, Image | Text | Proprietary | Jun 2026 |
| Gemini 4 Argon | $2 / $10 until Nov 30, 2026; $4 / $20 from Dec 1, 2026[8] | 1M | Not listed | Text | Proprietary | Not listed | |
| Gemini 3.8 Flash | $0.75 / $3.75 until Dec 31, 2026; $1.50 / $7.50 from Jan 1, 2027[9] | Not listed | Not listed | Text | Proprietary | Not listed | |
| Grok 4.7 | xAI | $2 / $6 (prompts under 200K tokens); $4 / $12 at 200K+[10] | 500K | Text, Image | Text | Proprietary | May 2026 |
| DeepSeek-V4-Pro | DeepSeek | $1.32 / $3.96 peak; half that off-peak[11] | 1M | Text | Text | Not listed | Not listed |
| Mistral Large 4 | Mistral AI | See vendor pricing[12] | Not listed | Text, Image | Text | Open weights | Not listed |
How we compile this table
Our full methodology covers column definitions, source tiers, the “Undisclosed” policy, recency policy, COI disclosure, and our correction process. Column structure intentionally mirrors Wikipedia's LLM comparison article for editor cross-reference.
Read the methodologyFrequently Asked Questions: LLM Comparisons
- What is the best large language model in 2026?
- There is no single best model; it depends on the job and the budget. OpenAI recommends GPT-6 Astra for complex reasoning and coding, and Anthropic recommends Claude Opus 5.5 for most workloads, with Claude Fable 5.1 for the most demanding reasoning. Google's flagship is Gemini 4 Argon and xAI's is Grok 4.7. For high-volume, cost-sensitive work, GPT-6 Luna and Claude Haiku 5.5 have the lowest list prices.
- Which LLM is cheapest to use through an API?
- Among current flagship-family models, GPT-6 Luna lists at $0.10 per million input tokens and $0.50 per million output tokens, and Claude Haiku 5.5 starts at the same rates. DeepSeek-V4-Pro lists at $1.32 / $3.96 at peak hours and half that off-peak. Batch processing and prompt caching cut these prices further on most platforms.
- Which LLM has the largest context window?
- OpenAI lists a 1.05M-token context window for GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna. Anthropic's current Claude models, Gemini 4 Argon and DeepSeek-V4-Pro list 1M tokens, and Grok 4.7 lists 500K. A context window is the total number of tokens a model can take in one request, including the prompt and the reply.
- Is Gemini 4 Argon getting more expensive?
- Yes. Google lists Gemini 4 Argon at $2 per million input tokens and $10 per million output tokens through November 30, 2026, rising to $4 and $20 from December 1, 2026. Gemini 4 Argon is not available on the Gemini API free tier.
- What is the difference between proprietary and open-weight LLMs?
- Proprietary models such as GPT-6, Claude and Gemini are only available through the vendor's apps and API, and their weights are not released. Open-weight models such as Mistral Large 4 publish their weights, so you can host, fine-tune and run them on your own hardware, subject to the model's license.
Sources
- GPT-6 Astra (OpenAI). OpenAI models (opens in new tab) Checked October 16, 2026.
- GPT-6.1 Sol (OpenAI). OpenAI models (opens in new tab) Checked October 16, 2026.
- GPT-6 Luna (OpenAI). OpenAI models (opens in new tab) Checked October 16, 2026.
- Claude Fable 5.1 (Anthropic). Anthropic models overview (opens in new tab) Checked October 16, 2026.
- Claude Opus 5.5 (Anthropic). Anthropic models overview (opens in new tab) Checked October 16, 2026.
- Claude Sonnet 5.5 (Anthropic). Anthropic models overview (opens in new tab) Checked October 16, 2026.
- Claude Haiku 5.5 (Anthropic). Anthropic models overview (opens in new tab) Checked October 16, 2026.
- Gemini 4 Argon (Google). Gemini API models (opens in new tab); Gemini API pricing (opens in new tab) Checked October 16, 2026.
- Gemini 3.8 Flash (Google). Gemini API models (opens in new tab); Gemini API pricing (opens in new tab) Checked October 16, 2026.
- Grok 4.7 (xAI). xAI models and pricing (opens in new tab) Checked October 16, 2026.
- DeepSeek-V4-Pro (DeepSeek). DeepSeek models and pricing (opens in new tab) Checked October 16, 2026.
- Mistral Large 4 (Mistral AI). Mistral models overview (opens in new tab) Checked October 16, 2026.
- * Updated Oct 2026: the lineup moved from GPT-4o, Claude 3.x, Gemini 2.x and Llama 3.3 to each vendor's current models, and an API list-price column replaced parameter counts, which the proprietary vendors do not publish. Data was checked against vendor model and pricing documentation on October 16, 2026. Prices and versions change often; vendor documentation is authoritative.