Compare / head-to-head
DeepSeek V4 Flashvs
GPT-6 Luna
DeepSeek V4 Flash leads 9 of 17 shared benchmarks. GPT-6 Luna has the larger context window (1.1M tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
9 – 8
DeepSeek V4 Flash leads
Cheaper per token
—
list price, input + output
Larger context
GPT-6 Luna
1.1M tokens
Providers
2 providers
DeepSeek · OpenAI
DeepSeek V4 Flash
DeepSeek · released 2026-07-31
text
- Context
- 1M
- Max out
- —
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 28 · 7 core
GPT-6 Luna
OpenAI · released 2026-09-22
textvision
Quality
Benchmark matrix
17 shared · 11 only DeepSeek V4 Flash · 5 only GPT-6 Luna| Benchmark | DeepSeek V4 Flash | GPT-6 Luna | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 17 | ||||
| GPQA Diamond | 91% ↗epoch run | 90.5% ↗epoch run | +0.5 pt | DeepSeek V4 Flash |
| LMArena Elo | 1436 ↗ | 1442.9 ↗ | -6.9 | GPT-6 Luna |
| Chess Puzzles (Epoch AI run) | 33% ↗epoch run | 31% ↗epoch run | +2 pt | DeepSeek V4 Flash |
| FrontierMath Tier 4 v2 (Epoch AI run) | 24.4% ↗epoch run | 56.1% ↗epoch run | -31.7 pt | GPT-6 Luna |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 57.5% ↗epoch run | 78.9% ↗epoch run | -21.4 pt | GPT-6 Luna |
| LiveBench Agentic Coding (LiveBench) | 46.8% ↗ | 51.2% ↗ | -4.4 pt | GPT-6 Luna |
| LiveBench Coding (LiveBench) | 75% ↗ | 79% ↗ | -4 pt | GPT-6 Luna |
| LiveBench Data Analysis (LiveBench) | 79.3% ↗ | 73.4% ↗ | +5.9 pt | DeepSeek V4 Flash |
| LiveBench Instruction Following (LiveBench) | 65.5% ↗ | 55.9% ↗ | +9.6 pt | DeepSeek V4 Flash |
| LiveBench Language (LiveBench) | 79.2% ↗ | 73.8% ↗ | +5.4 pt | DeepSeek V4 Flash |
| LiveBench Mathematics (LiveBench) | 86.8% ↗ | 89.1% ↗ | -2.3 pt | GPT-6 Luna |
| LiveBench Reasoning (LiveBench) | 86.6% ↗ | 81.8% ↗ | +4.8 pt | DeepSeek V4 Flash |
| LMArena WebDev (LMArena) | 1581 ↗ | 1578.7 ↗ | +2.3 | DeepSeek V4 Flash |
| Mystery Game Puzzles (Epoch AI run) | 34% ↗epoch run | 7% ↗epoch run | +27 pt | DeepSeek V4 Flash |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 94.4% ↗epoch run | 98.9% ↗epoch run | -4.5 pt | GPT-6 Luna |
| SimpleQA Verified | 33.6% ↗epoch run | 41.4% ↗epoch run | -7.8 pt | GPT-6 Luna |
| Terminal-Bench 4.0 (Vals AI) | 18.7% ↗ | 13.6% ↗ | +5.1 pt | DeepSeek V4 Flash |
| Only DeepSeek V4 Flash reports · 11 | ||||
| Agents' Last Exam | 25.2% ↗ | not reported | — | — |
| AutomationBench Public | 25.1% ↗ | not reported | — | — |
| Cybergym | 76.7% ↗ | not reported | — | — |
| DeepSWE | 54.4% ↗ | not reported | — | — |
| DeepSWE v1.1 (Datacurve) | 53.3% ↗ | not reported | — | — |
| NL2Repo | 54.2% ↗ | not reported | — | — |
| SimpleBench (SimpleBench) | 61.1% ↗ | not reported | — | — |
| Terminal Bench 2.1 | 82.7% ↗ | not reported | — | — |
| Toolathlon-Verified | 70.3% ↗ | not reported | — | — |
| Toolathlon-Verified (HKUST) | 70.7% ↗ | not reported | — | — |
| WeirdML (Håvard Tveit Ihle) | 63% ↗ | not reported | — | — |
| Only GPT-6 Luna reports · 5 | ||||
| APEX-Agents (Mercor) | not reported | 44.3% ↗ | — | — |
| DeepSWE v1.1 | not reported | 66.6% ↗ | — | — |
| Furniture Assembly (Epoch AI run) | not reported | 44.2% ↗epoch run | — | — |
| LMArena Agent (LMArena) | not reported | 0.0135 ↗ | — | — |
| SAGE (Vals AI) | not reported | 48.1% ↗ | — | — |
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 DeepSeek V4 Flash minus GPT-6 Luna 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 | DeepSeek V4 Flash | GPT-6 Luna | Edge |
|---|---|---|---|
| Context window | 1M tokens | 1.1M tokens | GPT-6 Luna |
| Max output | — | 128K tokens | — |
| Input price / 1M | — | $0.1 ↗ | — |
| Output price / 1M | — | $0.5 ↗ | — |
| Cached input / 1M | — | $0.01 ↗ | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | — | $0.6 | — |
| Modalities | text | text · vision | GPT-6 Luna |
| Released | 2026-07-31 | 2026-09-22 | — |
| Cited benchmark scores | 28 | 22 | — |
Reliability
Provider status
Live from /statusMore matchups
DeepSeek V4 Flash vs …
Models sharing the most benchmarksMore matchups
GPT-6 Luna vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run DeepSeek V4 Flash and GPT-6 Luna on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.