Qwen3.7 Max vs Ring-2.6-1T: Benchmarks, Pricing & Capabilities Compared
TL;DR — Qwen3.7 Max wins for reasoning + long-context · Ring-2.6-1T wins for cost.
Qwen3.7 Max Alibaba
- Released
- 2026-05-19
- Context window
- 1M tokens
- Input price
- $2.50 / Mtok
- Output price
- $7.50 / Mtok
Ring-2.6-1T InclusionAI
- Released
- 2026-05-08
- Context window
- 262K tokens
- Input price
- $0.30 / Mtok
- Output price
- $2.50 / Mtok
Benchmark comparison
| Benchmark | Qwen3.7 Max | Ring-2.6-1T |
|---|---|---|
| AA Intelligence Index | 56.6 ✓ | 38.5 |
| GPQA Diamond | 92.3% ✓ | 85.7% |
| HLE | 38.1% ✓ | 18.3% |
| IF-Bench | 80.5% ✓ | 44.6% |
| LiveCodeBench Reasoning | 69.0% ✓ | 64.3% |
| SciCode | 48.8% ✓ | 42.4% |
| TAU2-bench | 94.7% ✓ | 92.4% |
| TerminalBench-Hard | 50.8% ✓ | 28.8% |
Pricing comparison
| Metric | Qwen3.7 Max | Ring-2.6-1T |
|---|---|---|
| Input ($/Mtok) | $2.50 | $0.30 |
| Output ($/Mtok) | $7.50 | $2.50 |
| Cached input ($/Mtok) | — | — |
| Cost per 1M-token roundtrip (1M in + 1M out) | $10.00 | $2.80 |
Context window & modalities
| Attribute | Qwen3.7 Max | Ring-2.6-1T |
|---|---|---|
| Context window | 1M tokens | 262K tokens |
| Input modalities | text | text |
| Output modalities | text | text |
| Knowledge cutoff | — | — |
Verdict by use case
Coding
Insufficient data
Basis: SWE-bench
No shared coding benchmark.
Reasoning
→ Qwen3.7 Max
Basis: GPQA Diamond
Qwen3.7 Max 92.3% vs Ring-2.6-1T 85.7% on GPQA Diamond.
Math
Insufficient data
Basis: MATH / AIME
No shared math benchmark.
Long context
→ Qwen3.7 Max
Basis: Context window
Qwen3.7 Max 1M tokens vs Ring-2.6-1T 262K tokens.
Cost
→ Ring-2.6-1T
Basis: Input $/Mtok
Qwen3.7 Max $2.5/Mtok vs Ring-2.6-1T $0.3/Mtok input.
Changelog & releases
Qwen3.7 Max
Released 2026-05-19
Ring-2.6-1T
Released 2026-05-08