Gemini vs Sol
Research multimancer vs coding ceiling — when is 96% SWE worth 2.5× the output tokens? Public stats are peer-relative ranks (40–95 band among scored Builds), not absolute “% intelligence.”
Gemini 3.1 Pro · API · High thinking
The reasoning multimancer — 1M context, strong science, calibrated knowledge.
GPT-5.6 Sol · API · Max effort
The ecosystem operator — max-effort Sol for hard agents, not cheap chat.
Six attributes (peer public)
Same-harness & loadout
| Signal | Gemini | Sol |
|---|---|---|
| SWE Verified | 79.8% openlm · 80.6% vendor | 96.2% Vals mini-swe |
| GPQA Diamond | 94.1% AA (tied) | 94.1% AA (tied) |
| HLE (closed-book) | 44.4% vendor | 47.2% AA |
| Omniscience / calibration | Index ~33 (leader narrative) | High acc · high hall (WIS peer-weak) |
| List price (in/out ≤200k) | $2 / $12 | $5 / $30 |
| MMMU-Pro (vision+reasoning) | 82% AA · 80.5% vendor | — (not pinned this Build) |
| Modalities | Text · image · audio · video | Text · image |
Deep Research
Decisive on Judgment: Gemini — public ranges do not overlap (Gemini 91.1–99.1 vs Sol 35.9–43.9).
- Deep ResearchGemini · conf A
Co-leading GPQA, stronger calibration narrative, 1M multimodal context.
Quest Fit → - Large Existing CodebaseSol · conf A
Independent mini-swe gap is large; peer STR also Sol.
Quest Fit → - Autonomous Coding AgentSol · conf A
Sol leads agentic boards and long-horizon terminal work.
Quest Fit → - Polished FrontendGemini · conf A
AA MMMU-Pro ~82% + native multimodal + vision Arena; academic multimodal ≠ pure craft.
Quest Fit → - Under BudgetGemini · conf A
Roughly half Sol’s list output price; still not a Flash sprinter.
Quest Fit →
Gemini: 1M-context multimodal Pro with co-leading GPQA and a peer-leading Omniscience story at $2/$12.
Sol: Independent SWE Verified peak of the peer set (96.2% mini-swe) — pay $5/$30 and max-effort tax.