Inference
EN

Seed 2.1 Pro vs Seed 2.1 Turbo

Change models
Seed 2.1 ProByteDance
Seed 2.1 TurboByteDance
IndexAll categories, out of 100.
60.1 ±8.9 best
57.4 ±8.9
GapIndex points below the top.
Top
−2.7 · no clear gap
CoverageIndex weight that has results.
65%
65%
SpeedOutput tokens per second.
—
37/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-06
2026-06
Input / 1MUS dollars per 1M input tokens.
—
$0.50
Output / 1MUS dollars per 1M output tokens.
—
$2.50
1K chats2K tokens in, 500 out, each.
—
$2.25
1K documents20K tokens in, 1K out, each.
—
$12.50
1K agent tasks200K tokens in, 20K out, each.
—
$150.00
ContextMaximum tokens in one request.
256K
262K best
AgenticMulti-step tasks with tools.
62.7 best
60.7
CodingCode writing and repair.
59.9 best
57.6
ReasoningLogic problems and puzzles.
—
—
MultimodalTasks with images and text.
62.7 best
59.3
KnowledgeFacts and expert knowledge.
55.2 best
52.4
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
—
—

Benchmarks

10 shared
Seed 2.1 Pro
Seed 2.1 Turbo
Terminal-Bench 2.1 · BenchLMAgentic
71.0% best49 / 86
67.6%54 / 86
Massive Multi-discipline Multimodal Understanding ProMultimodal
81.6% best7 / 53
80.1%15 / 53
CharXiv ReasoningMultimodal
85.4% best19 / 49
82.5%23 / 49
NL2RepoCoding
47.0% best18 / 36
43.7%20 / 36
MathVisionMultimodal
92.6% best4 / 24
90.1%8 / 24
SuperGPQA: Scaling LLM Evaluation Across 285 Graduate DisciplinesKnowledge
70.8% best7 / 22
67.4%11 / 22
ERQAMultimodal
72.0% best3 / 18
71.3%4 / 18
BabyVisionMultimodal
73.7% best4 / 11
62.9%7 / 11
ZeroBenchMultimodal
18.0% best9 / 10
11.0%10 / 10
Video-MMEMultimodal
89.2%
89.0%