Compare / head-to-head
Gemini 4 Argonvs
Kimi K2.5
Gemini 4 Argon leads 4 of 4 shared benchmarks.
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
4 – 0
Gemini 4 Argon leads
Cheaper per token
—
list price, input + output
Larger context
—
unknown
Providers
2 providers
Google · Moonshot AI
Gemini 4 Argon
Google · released 2026-09-30
textvisionvideo
Kimi K2.5
Moonshot AI · released 2026-01-27
textvisionvideo
- Context
- 256K
- Max out
- —
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 26 · 14 core
Quality
Benchmark matrix
4 shared · 8 only Gemini 4 Argon · 22 only Kimi K2.5| Benchmark | Gemini 4 Argon | Kimi K2.5 | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 4 | ||||
| LMArena Elo | 1525.2 ↗ | 1450.1 ↗ | +75.1 | Gemini 4 Argon |
| LMArena WebDev (LMArena) | 1679.5 ↗ | 1435.8 ↗ | +243.7 | Gemini 4 Argon |
| SAGE (Vals AI) | 53.6% ↗ | 49.9% ↗ | +3.7 pt | Gemini 4 Argon |
| Vending-Bench 2 (Andon Labs) | 13718.16 ↗ | 1198.46 ↗ | +12519.7 | Gemini 4 Argon |
| Only Gemini 4 Argon reports · 8 | ||||
| APEX-Agents (Mercor) | 82.2% ↗ | not reported | — | — |
| AutomationBench | 51.3% ↗ | not reported | — | — |
| CWE-bench v1 | 68% ↗ | not reported | — | — |
| DeepSWE v1.1 | 77.9% ↗ | not reported | — | — |
| FrontierSWE V2 (Proximal Labs) | 55% ↗ | not reported | — | — |
| LMArena Agent (LMArena) | 0.0757 ↗ | not reported | — | — |
| LVBench | 91.7% ↗ | not reported | — | — |
| Terminal-Bench 4.0 (Vals AI) | 57.6% ↗ | not reported | — | — |
| Only Kimi K2.5 reports · 22 | ||||
| SWE-bench Verified | not reported | 76.8% ↗ | — | — |
| MMLU-Pro | not reported | 87.1% ↗ | — | — |
| AIME 2025 | not reported | 96.1% ↗ | — | — |
| BrowseComp | not reported | 60.6% ↗ | — | — |
| GPQA-Diamond | not reported | 87.6% ↗ | — | — |
| HLE-Full | not reported | 30.1% ↗ | — | — |
| HLE-Full (w/ tools) | not reported | 50.2% ↗ | — | — |
| HMMT 2025 (Feb) | not reported | 95.4% ↗ | — | — |
| IMO-AnswerBench | not reported | 81.8% ↗ | — | — |
| LiveCodeBench (v6) | not reported | 85% ↗ | — | — |
| LMArena Vision (LMArena) | not reported | 1252.2 ↗ | — | — |
| MMMU-Pro | not reported | 78.5% ↗ | — | — |
| OSWorld-Verified (XLANG) | not reported | 63.3% ↗ | — | — |
| SimpleBench (SimpleBench) | not reported | 46.8% ↗ | — | — |
| SimpleQA Verified | not reported | 34.3% ↗epoch run | — | — |
| SWE-Bench Multilingual | not reported | 73% ↗ | — | — |
| SWE-Bench Pro | not reported | 50.7% ↗ | — | — |
| SWE-bench Verified (Epoch AI run) | not reported | 73.8% ↗epoch run | — | — |
| Terminal Bench 2.0 | not reported | 50.8% ↗ | — | — |
| Toolathlon-Verified (HKUST) | not reported | 33% ↗ | — | — |
| VideoMMMU | not reported | 86.6% ↗ | — | — |
| WeirdML (Håvard Tveit Ihle) | not reported | 45.6% ↗ | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Gemini 4 Argon minus Kimi K2.5 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | Gemini 4 Argon | Kimi K2.5 | Edge |
|---|---|---|---|
| Context window | — | 256K tokens | — |
| Max output | 1M tokens | — | — |
| Input price / 1M | $2 ↗ | — | — |
| Output price / 1M | $10 ↗ | — | — |
| Cached input / 1M | $0.1 ↗ | — | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $12 | — | — |
| Modalities | text · vision · video | text · vision · video | Tie |
| Released | 2026-09-30 | 2026-01-27 | — |
| Cited benchmark scores | 12 | 26 | — |
Reliability
Provider status
Live from /statusMore matchups
Gemini 4 Argon vs …
Models sharing the most benchmarksMore matchups
Kimi K2.5 vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Gemini 4 Argon and Kimi K2.5 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.