Ling 2.6 1T vs Ling 3.1 Flash
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Ling 2.6 1T8.7
Ling 3.1 Flash39.4
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Ling 2.6 1T—
Ling 3.1 Flash33.3
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Ling 2.6 1T—
Ling 3.1 Flash61.7
| Benchmark | Ling 2.6 1T | Ling 3.1 Flash |
|---|---|---|
| Humanity's Last Exam | 8.7 | 39.4 |
| Terminal-Bench 4.0 | — | 33.3 |
| AutomationBench-AA | — | 61.7 |
Specs
| Spec | Ling 2.6 1T | Ling 3.1 Flash |
|---|---|---|
| Input, per 1M | — | $0.30 |
| Output, per 1M | — | $0.90 |
| Cost per task? | — | $1.00 |
| Speed, tok/s | — | 145 |
| Context | 262K | 1M |
| Weights | Open | Open |
| Released | Apr 2026 | Oct 2026 |
| Memory (4-bit) | 615 GB | 343 GB |
| Licence | MIT | MIT |
| Commercial use | Allowed | Allowed |
| See images | — | — |
| Read PDFs | — | — |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | ? | Partial |
| Return strict JSON | Partial | ✓ |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Ling 2.6 1T → Ling 3.1 Flash
Nothing known.
Ling 3.1 Flash → Ling 2.6 1T