Skip to content
Model Pareto

GPT-6 Astra vs Gemini 4 Argon

Gemini 4 Argon scores higher than GPT-6 Astra in all 3 shared categories. GPT-6 Astra costs 5× as much as Gemini 4 Argon ($20.0 vs $4.00 per million tokens, blended 3:1 input:output).
GPT-6 AstraGemini 4 Argon
Focus

Coding: quality vs price

Compared models are ringed; the other 135 ranked here are greyed.

Lower price is better. Pareto frontier: Granite 4.2 3B, NVIDIA Nemotron 3 Nano 30B A3B, Nemotron 3.5 Lightning, gpt-oss-20b, Ling-3.0-flash-VL, MiMo-V2.6-Flash, GLM-5.3 Flash, MiniMax-M3.1-Flash, DeepSeek V4.1 Flash 1017, MiMo-V2.6-Pro, Kimi K3.1 Pro, Claude Opus 5.5, Claude Fable 5.2. 43 models are not plotted: Muse Wonder Full, GLM-5.3 Flash NVFP4, DeepSeek V4.1 Flash NVFP4, DeepSeek-V4-Pro-0813-nvfp4-DSpark, GLM-5.3 NVFP4, Xing4.0-29B-A4B, Ternary Bonsai 27B, Atria Dawn Preview, Maple-Preview, SWE-2 Flash, Ripple-65B Preview, Beam 1.0 501B A23B, Qwen3.8 27B NVFP4, Intern-S2-397B, Bonsai 27B, Solar Open2 250B, Muse Spark Open, Phi-4-reasoning-plus, Laguna S 2.1, Flint, Nex-N2-Pro, Motif 3, Nemotron 3.5 Super VL NVFP4, Nemotron 3.5 Super VL, K2 Horizon 375B A23B, K-EXAONE 2.0 0803, A.X-K2, Nemotron 3.5 Bolt 74B-A9B, K2 Horizon MoVA 36B A4B, Intern-S2-Preview (35B-A3B), North Mini Code, G9v3-39A5B, LFM2.5-24B-A2B, K2 Horizon 7B, Qwen3.8 Flash-Next NVFP4, K2-Horizon-32B, NVIDIA Nemotron-Labs Dory1.0 4B, Falcon-H1-34B-Instruct, MiniCPM5-2B, Ling 3.0 Tiny, K2 Horizon 3.7B, K2 Horizon 0.9B, LFM2.5-2.6B.

43 models with no price data — shown in the strip at the left edge

Muse Wonder Full, GLM-5.3 Flash NVFP4, DeepSeek V4.1 Flash NVFP4, DeepSeek-V4-Pro-0813-nvfp4-DSpark, GLM-5.3 NVFP4, Xing4.0-29B-A4B, Ternary Bonsai 27B, Atria Dawn Preview, Maple-Preview, SWE-2 Flash, Ripple-65B Preview, Beam 1.0 501B A23B, Qwen3.8 27B NVFP4, Intern-S2-397B, Bonsai 27B, Solar Open2 250B, Muse Spark Open, Phi-4-reasoning-plus, Laguna S 2.1, Flint, Nex-N2-Pro, Motif 3, Nemotron 3.5 Super VL NVFP4, Nemotron 3.5 Super VL, K2 Horizon 375B A23B, K-EXAONE 2.0 0803, A.X-K2, Nemotron 3.5 Bolt 74B-A9B, K2 Horizon MoVA 36B A4B, Intern-S2-Preview (35B-A3B), North Mini Code, G9v3-39A5B, LFM2.5-24B-A2B, K2 Horizon 7B, Qwen3.8 Flash-Next NVFP4, K2-Horizon-32B, NVIDIA Nemotron-Labs Dory1.0 4B, Falcon-H1-34B-Instruct, MiniCPM5-2B, Ling 3.0 Tiny, K2 Horizon 3.7B, K2 Horizon 0.9B, LFM2.5-2.6B

  • Best-value frontier (nothing is both cheaper and better) · strong + medium evidence; weak shown but excluded
  • Evidencestrong → weak
  • Estimated from other categories
  • GPT-6 Astra
    Quality
    83.4
    Rank
    #11/137
    Price
    $20.0/M tok
    Speed
    52tok/s
    In$10Out$50/M tok

    vs Gemini 4 Argon: −4.7 quality · 5× price

  • Gemini 4 Argon
    Quality
    88.1
    Rank
    #6/137
    Price
    $4.00/M tok
    Speed
    —
    In$2.00▾Out$10▾/M tok

    vs GPT-6 Astra: +4.8 quality · 0.2× price

Overview

Lab
OpenAI
Google DeepMind
Released
Sep 3, 2026
Sep 30, 2026
Weights
Proprietary
Proprietary
Input price
$/M tok
$10
$2.00Best
Output price
$/M tok
$50
$10Best
Blended price
$/M tok · 3:1 in:out
$20
$4.00Best
Output speed
tok/s
52
—
Context
tokens
1MBest
1MBest

Quality by category

0–100 within each category (100 = best tracked model). Rank is among all models ranked there. Click a row to focus it.

83.4±4.5
#11/137Verified
88.1±4.5Best
#6/137Verified
89.3±5.3
#3/162Verified
94.4±6.0Best
#2/162Verifiedbest value
97.2±3.8
#1/98Verifiedbest value
Not ranked
87.6±3.7
#8/218Verified
87.9±5.1Best
#7/218Verified
80.6±4.6
#16/187Mixed
Not ranked

Coding benchmarks

Rank among models with a published score, and the leaderboard around each model. ~ italic = no published score, estimated from related benchmarks.

~#2/21~89.1%Estimated
estimated from related benchmarks
~#2/21~93.2%Estimated
estimated from related benchmarks
Benchmarks in other categories (4)

Writing & chat benchmarks

Rank among models with a published score, and the leaderboard around each model. ~ italic = no published score, estimated from related benchmarks.

Vision benchmarks

Rank among models with a published score, and the leaderboard around each model. ~ italic = no published score, estimated from related benchmarks.

Hard reasoning benchmarks

Rank among models with a published score, and the leaderboard around each model. ~ italic = no published score, estimated from related benchmarks.

Agents benchmarks

Rank among models with a published score, and the leaderboard around each model. ~ italic = no published score, estimated from related benchmarks.