Gemini 4 Argon vs Claude Sonnet 5.5 vs GPT-6.1 Sol vs Opus 5.5 vs GPT-6 Astra

Gemini 4 Argon is Google's new flagship model for real-world coding, enterprise knowledge work and cyber defense. It takes text, image, audio, video and PDF input. Its 1M-token context is shared by input, thinking and output, and responses can run past 64K tokens.

Priced like the mid tier, benchmarked against the top tier. Through November 30, 2026, Argon costs an introductory $2 / $10 per 1M tokens (cached input $0.10), the same as Claude Sonnet 5.5 and GPT-6.1 Sol. Google compares it with Opus 5.5 ($4 / $20) and GPT-6 Astra ($10 / $50). After the introductory period the standard price is $4 / $20.

From Google's model card (self-reported):

• First or tied on 14 of 19 rows against Astra, Opus 5.5 and Fable 5.1
• DeepSWE v1.1 77.9 vs 74.2 (Opus 5.5) and 74.1 (Astra)
• Long context: GraphWalks BFS 256k–1M 84.2 vs 71.8 (Astra) and 66.8 (Opus 5.5)
• Knowledge work: Harvey Legal Agent 19.6 vs 6.7 or lower for the others, Finance Agent v2 65.4 vs 58.6 (Opus 5.5)
• Where it trails: Terminal-bench 4.0 (57.4 vs 66.4 for Opus 5.5) and FrontierSWE v2 (55.0 vs 65.5 for Astra)

Independent view (Artificial Analysis Intelligence Index): Argon (High) scores 52.6, level with GPT-6 Astra (52.7) and ahead of GPT-6.1 Sol (51.8). Sonnet 5.5 (56.0) and Opus 5.5 (57.6) still lead at max effort. Argon cost about $1.99 per index task, against $5.46 for Sonnet 5.5 and $5.98 for Opus 5.5. GPT-6.1 Sol is cheaper still at $0.72.

Gemini 4 Argon announcement · Model card