Qwen3.8-Max vs Claude Fable 5.1

Claude Fable 5.1 leads on the Visual Intelligence Index by 2.4 points, and Qwen3.8-Max scores higher on 3 of the 7 benchmarks where the two differ.

Model

Compared with

At a glance

Measure	Qwen3.8-Max	Claude Fable 5.1
Visual Intelligence Index	73.8%	76.2%
Rank	#5 of 10	#4 of 10
Cost per taskMean over the benchmarks, at list price.	$0.151	$0.517
Latency per taskEach benchmark's median, averaged.	66.0s	9.0s
Tokens per task	68,932	51,151
Input / 1M tokens	$2.00	$10.00
Output / 1M tokens	$6.00	$50.00
Context window	1M	1M
Availability	Proprietary	Proprietary

Each capability index is a weighted mean of the benchmarks that contribute to the capability.

Capability	Qwen3.8-Max	Claude Fable 5.1	Lead
Perception	72.0%	75.3%	−3.3
Spatial Reasoning	74.3%	77.0%	−2.8
Temporal Reasoning	73.6%	76.4%	−2.8
Action / Event Understanding	71.5%	74.0%	−2.4
Causal Reasoning	78.2%	78.7%	−0.5
Information Retrieval	72.0%	74.7%	−2.7

Benchmark	Qwen3.8-Max	Claude Fable 5.1	Lead
Video-MME v2	60.2%	66.2%	−6.0
LVBench	58.9%	75.0%	−16.1
Perception Test	83.9%	77.3%	+6.6
NExT-QA	87.6%	86.8%	+0.8
Q-Bench Video	69.0%	71.0%	−2.0
EgoSchema	85.2%	80.4%	+4.8
UCF101-AD	56.5%	60.6%	−4.1

3.5/ 5

Reachable1.0: There is an endpoint you can call without hosting anything.
First-party API — supported
Third-party API — supported
Portable0.0: You can run it yourself, and are not tied to one vendor.
Open weights — not supported
Self-hostable — not supported
Multimodal input0.5: It takes the footage directly, rather than frames you extracted.
Video ingestion — supported
Audio — not supported
Programmable1.0: Output you can parse, and tools it can call on its own.
Structured output — supported
Tool calling — supported
Operable1.0: Usable interactively and in bulk, not only one call at a time.
Streaming — supported
Batch API — supported

3.0/ 5

Reachable1.0: There is an endpoint you can call without hosting anything.
First-party API — supported
Third-party API — supported
Portable0.0: You can run it yourself, and are not tied to one vendor.
Open weights — not supported
Self-hostable — not supported
Multimodal input0.0: It takes the footage directly, rather than frames you extracted.
Video ingestion — not supported
Audio — not supported
Programmable1.0: Output you can parse, and tools it can call on its own.
Structured output — supported
Tool calling — supported
Operable1.0: Usable interactively and in bulk, not only one call at a time.
Streaming — supported
Batch API — supported