gemini-4-argon
96.4%±1.2%
95.2%CI97.6%
aiqbench · LLM benchmarks
Benchmarks
- Sources
- 8
- Entries tracked
- 22,310
- Last sync
- 33 min33 minutes ago
All sources syncing normally
- Arena · Text· 2 hours ago
- Arena · WebDev· 2 hours ago
- Arena · Vision· 2 hours ago
- Arena · Agent· 2 hours ago
- GPQA Diamond· 33 minutes ago
- SWE-bench· 33 minutes ago
- Terminal-Bench· 33 minutes ago
- ARC-AGI-2· 33 minutes ago
Filter by org:
Released:
gpt-6-astra
95.8%±1.4%
94.4%CI97.1%
claude-sonnet-5-5
95.6%±1.4%
94.2%CI97.0%
One grid, every column. Pick axes in the Axes panel — one axis ranks like a single board, several blend into a weighted composite, all of them read as the index. Rank is assigned before any filter and never renumbers.
| Model | Org | ± CI | Swing | Add to compare | |||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | gemini-4-argon | Google DeepMind | Oct 2026 | $2 / $10 | 96.4% | 95.2%–97.6% | #1–10 | ||
| 2 | gpt-6-astra | OpenAI | Sep 2026 | $10 / $50 | 95.8% | 94.4%–97.1% | #1–18 | ||
| 3 | claude-sonnet-5-5 | Anthropic | Oct 2026 | $2 / $10 | 95.6% | 94.2%–97.0% | #4–26 | ||
| 4 | gemini-3.8-flash | Google DeepMind | Oct 2026 | $0.83 / $4.13 | 95.4% | 94.0%–96.8% | #2–29 | ||
| 5 | gpt-6.1-sol | OpenAI | Oct 2026 | $2 / $10 | 95.4% | 94.0%–96.8% | #1–15 | ||
| 6 | gemini-3.7-flash | Google DeepMind | Oct 2026 | $0.75 / $3.75 | 94.8% | 93.5%–96.2% | #9–57 | ||
| 7 | gpt-5.4-pro-2026-03-05 | OpenAI | Mar 2026 | $30 / $180 | 94.6% | 93.0%–96.2% | #20–171 | ||
| 8 | gemini-3.1-pro-preview | Google DeepMind | Sep 2026 | $2 / $12 | 94.4% | 92.8%–96.1% | #9–36 | ||
| 9 | gpt-6-sol | OpenAI | Oct 2026 | $2 / $10 | 94.3% | 92.8%–95.8% | #7–53 | ||
| 10 | gemini-3.6-flash | Google DeepMind | Sep 2026 | $0.75 / $3.75 | 94.1% | 92.7%–95.5% | #16–95 | ||
| 11 | gpt-5.5-pre-release | OpenAI | — | — | 94.0% | 92.5%–95.5% | #22–36 | ||
| 12 | grok-4.6 | xAI | Sep 2026 | $2 / $6 | 94.0% | 92.6%–95.5% | #15–52 | ||
| 13 | gpt-5.5-pro-pre-release | OpenAI | — | — | 93.9% | 92.3%–95.5% | #23–60 | ||
| 14 | claude-opus-5 | Anthropic | Sep 2026 | $5 / $25 | 93.9% | 92.4%–95.4% | #5–11 | ||
| 15 | gpt-5.6-sol | OpenAI | Sep 2026 | $4 / $20 | 93.5% | 91.9%–95.1% | #6–25 | ||
| 16 | grok-4.5 | xAI | Sep 2026 | $2 / $6 | 93.4% | 92.0%–94.9% | #29–70 | ||
| 17 | gpt-5.6-terra | OpenAI | Sep 2026 | $2 / $12 | 93.3% | 91.8%–94.8% | #13–59 | ||
| 18 | gpt-5.4-2026-03-05 | OpenAI | Sep 2026 | $2.50 / $15 | 93.3% | 91.5%–95.1% | #12–29 | ||
| 19 | kimi-k3 | Moonshot | Jun 2026 | $0.25 / $11.94 | 93.1% | 91.6%–94.6% | #10–39 | ||
| 20 | gemini-3.5-flash | Google DeepMind | Sep 2026 | $1.50 / $9 | 92.8% | 91.2%–94.4% | #10–26 | ||
| 21 | grok-4.7 | xAI | Oct 2026 | $2 / $6 | 92.7% | 91.2%–94.3% | #21–72 | ||
| 22 | qwen3.8-max | Alibaba | Oct 2026 | $2 / $6 | 92.7% | 91.0%–94.4% | #12–40 | ||
| 23 | gemini-3-pro-preview | Google DeepMind | — | — | 92.6% | 91.0%–94.3% | #19–120 | ||
| 24 | qwen3.8-max-0902 | Alibaba | Oct 2026 | $2 / $6 | 92.3% | 90.6%–94.0% | #23–155 | ||
| 25 | glm-5.2 | Z.ai (Zhipu AI) | Jun 2026 | $0.62 / $2.42 | 91.9% | 90.2%–93.5% | #9–49 | ||
| 26 | deepseek-v4-pro-0813 | DeepSeek | Aug 2026 | $0.66 / $1.98 | 91.7% | 90.1%–93.2% | #40–130 | ||
| 27 | gpt-5.6-luna | OpenAI | Sep 2026 | $0.20 / $1.20 | 91.6% | 89.9%–93.3% | #30–146 | ||
| 28 | gpt-5.2-2025-12-11 | OpenAI | Dec 2025 | $1.75 / $14 | 91.4% | 89.6%–93.2% | #34–92 | ||
| 29 | claude-opus-4-8 | Anthropic | Sep 2026 | $5 / $25 | 91.0% | 89.1%–93.0% | #15–33 | ||
| 30 | deepseek-v4-flash-0731 | DeepSeek | Jul 2026 | $0.02 / $0.45 | 91.0% | 89.3%–92.8% | #45–244 | ||
| 31 | deepseek-v4-pro | DeepSeek | Aug 2026 | $0.66 / $1.98 | 90.9% | 88.9%–93.0% | #17–54 | ||
| 32 | glm-5.3 | Z.ai (Zhipu AI) | Aug 2026 | $0.08 / $1 | 90.9% | 89.3%–92.5% | #18–52 | ||
| 33 | MiniMax-M3 | MiniMax | Jun 2026 | $0.30 / $1.20 | 90.9% | 88.9%–93.0% | #50–264 | ||
| 34 | qwen3.7-max | Alibaba | May 2026 | $1.48 / $4.42 | 90.9% | 88.9%–93.0% | #41–128 | ||
| 35 | kimi-k2.6 | Moonshot | Apr 2026 | $0.65 / $3.41 | 90.8% | 89.1%–92.5% | #29–78 |
Showing GPQA Diamond exactly as the source published it — its own score and confidence interval, no blending. Select a second axis to build a weighted composite.