Compare / head-to-head
Llama 3.3 70B InstructvsQwen3.5 397B-A17B
Qwen3.5 397B-A17B leads 4 of 4 shared benchmarks. Qwen3.5 397B-A17B has the larger context window (262K tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 4
Qwen3.5 397B-A17B leads
Cheaper per token
—
list price, input + output
Larger context
Qwen3.5 397B-A17B
262K tokens
Provider uptime (30d)
— · —
Meta · Qwen
Llama 3.3 70B Instruct
Meta · released 2024-12-06
text
- Context
- 131K
- Max out
- —
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 5 · 4 core
Qwen3.5 397B-A17B
Qwen · released 2026-02-15
textvision
Quality
Benchmark matrix
4 shared · 0 only Llama 3.3 70B Instruct · 3 only Qwen3.5 397B-A17B| Benchmark | Llama 3.3 70B Instruct | Qwen3.5 397B-A17B | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 4 | ||||
| GPQA Diamond | 50.5% ↗ | 88.4% ↗ | -37.9 pt | Qwen3.5 397B-A17B |
| MMLU-Pro | 68.9% ↗ | 87.8% ↗ | -18.9 pt | Qwen3.5 397B-A17B |
| LMArena Elo | 1274 ↗ | 1438.3 ↗ | -164.3 | Qwen3.5 397B-A17B |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 5.1% ↗epoch | 88.9% ↗epoch | -83.8 pt | Qwen3.5 397B-A17B |
| Only Llama 3.3 70B Instruct reports · 0 | ||||
| Llama 3.3 70B Instruct reports nothing Qwen3.5 397B-A17B does not. | ||||
| Only Qwen3.5 397B-A17B reports · 3 | ||||
| SWE-bench Verified | not reported | 76.4% ↗ | — | — |
| AIME 2026 | not reported | 94.2% ↗matharena ⚠ | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | not reported | 29.5% ↗epoch | — | — |
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 Llama 3.3 70B Instruct minus Qwen3.5 397B-A17B 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 | Llama 3.3 70B Instruct | Qwen3.5 397B-A17B | Edge |
|---|---|---|---|
| Context window | 131K tokens | 262K tokens | Qwen3.5 397B-A17B |
| Max output | — | — | — |
| Input price / 1M | — | $0.172 ↗ | — |
| Output price / 1M | — | $1.03 ↗ | — |
| Cached input / 1M | — | — | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | — | $1.2 | — |
| Modalities | text | text · vision | Qwen3.5 397B-A17B |
| Released | 2024-12-06 | 2026-02-15 | — |
| Cited benchmark scores | 5 | 8 | — |
Reliability
Provider status
Last 60 days · refreshed daily on this page · live on /statusMore matchups
Llama 3.3 70B Instruct vs …
Models sharing the most benchmarksMore matchups
Qwen3.5 397B-A17B vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Llama 3.3 70B Instruct and Qwen3.5 397B-A17B on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.