someaistuff.
Side by side

See the difference.

Scores, capabilities and costs for the models you’re considering.

Scroll to compare models →

Comparison of Step-5-Preview, Xiaomi: MiMo-V2.6-Pro, Reflection: Beam 1.0 501B A23B
3 modelsstepfun-aiStep-5-PreviewRemove ×xiaomiXiaomi: MiMo-V2.6-ProRemove ×reflectionReflection: Beam 1.0 501B A23BRemove ×
What’s different600B total and 27B active parameters; accepts images and video, with FP8 and BF16 weights for deployment.MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi. Built at a scale of over 1T parameters, it is designed to push the ceiling of capability for the most demanding...Apache-2.0 text model with 501B total and 23B active parameters; ships BF16, FP8 and NVFP4 weights.
DateSep 18, 2026ReleasedSep 21, 2026ReleasedOct 16, 2026Released
API prices USD · provider configuration shown below
Listed rate$1.00 in / $2.70 out/ 1M tokensDeepInfra · all routes ↗$0.387 in / $0.783 out/ 1M tokensDeepInfra · via OpenRouter · all routes ↗deepinfra/fp8 · quantization=fp8$0.25 in / $1.00 out/ 1M tokensNebius · via OpenRouter · all routes ↗nebius/fp8 · quantization=fp8
Provider endpoints6 endpoints →6 endpoints →6 endpoints →
Benchmarks Each result retains its evaluation configuration
AA Intelligence ↗4.3.243.73 ↗Effort not specifiedIndependent result46.32 ↗Effort not specifiedIndependent result30.33 ↗Effort not specifiedIndependent result
AA-Briefcase ↗1.11,424.32 ↗Effort not specifiedIndependent result1,513.98 ↗Effort not specifiedIndependent result976.2 ↗Effort not specifiedIndependent result
GDPval-AA ↗2.11,586.63 ↗Effort not specifiedIndependent result1,683.29 ↗Effort not specifiedIndependent result1,177.08 ↗Effort not specifiedIndependent result
AutomationBench-AA ↗AA Index 4.3.2 collection50.96% ↗Effort not specifiedIndependent result58.62% ↗Effort not specifiedIndependent result48.46% ↗Effort not specifiedIndependent result
Terminal-Bench 4.0 · AA ↗33.33% ↗Effort not specifiedHarness unreportedIndependent result34.85% ↗Effort not specifiedHarness unreportedIndependent result3.03% ↗Effort not specifiedHarness unreportedIndependent result
SciCode ↗AA Index 4.3.2 collection58.91% ↗AA · Under reviewEffort not specifiedIndependent result60.88% ↗AA · Under reviewEffort not specifiedIndependent result48.5% ↗AA · Under reviewEffort not specifiedIndependent result
Humanity’s Last Exam ↗AA Index 4.3.2 collection46.48% ↗Effort not specifiedIndependent result49.35% ↗Effort not specifiedIndependent result34.66% ↗Effort not specifiedIndependent result
GDP.pdf ↗AA Index 4.3.2 collection14.8% ↗Effort not specifiedIndependent result19.2% ↗Effort not specifiedIndependent result11.6% ↗Effort not specifiedIndependent result
CritPt ↗AA Index 4.3.2 collection20.86% ↗AA · Under reviewEffort not specifiedIndependent result26.57% ↗AA · Under reviewEffort not specifiedIndependent result13.14% ↗AA · Under reviewEffort not specifiedIndependent result
AA-Omniscience ↗AA Index 4.3.2 collection16.38 ↗Effort not specifiedIndependent result8.38 ↗Effort not specifiedIndependent result1.5 ↗Effort not specifiedIndependent result
AA-LCR ↗1.188.33% ↗Effort not specifiedIndependent result86.33% ↗Effort not specifiedIndependent result78% ↗Effort not specifiedIndependent result
LiveBench ↗2.0 · 2026-10-14 (dataset 2026-09-02)62.01% ↗Effort: max · single run · equal mean of 7 suitesIndependent result74.87% ↗Effort: high · single run · equal mean of 7 suitesIndependent result55.5% ↗Effort: max · single run · equal mean of 7 suitesIndependent result
LiveBench Design ↗2.0 · 2026-10-14 (dataset 2026-09-02)76.9% ↗Effort: max · single runIndependent result78.5% ↗Effort: high · single runIndependent result67.9% ↗Effort: max · single runIndependent result
LiveBench Research ↗2.0 · 2026-10-14 (dataset 2026-09-02)72.8% ↗Effort: max · single runIndependent result76.7% ↗Effort: high · single runIndependent result72.8% ↗Effort: max · single runIndependent result
LiveBench Browser Use ↗2.0 · 2026-10-14 (dataset 2026-09-02)24% ↗Effort: max · single runIndependent result88% ↗Effort: high · single runIndependent result0% ↗Effort: max · single runIndependent result
LiveBench Coding ↗2.0 · 2026-10-14 (dataset 2026-09-02)85.6% ↗Effort: max · single runIndependent result81.5% ↗Effort: high · single runIndependent result83.9% ↗Effort: max · single runIndependent result
LiveBench Agentic ↗2.0 · 2026-10-14 (dataset 2026-09-02)27.5% ↗Effort: max · single runIndependent result47.5% ↗Effort: high · single runIndependent result17.3% ↗Effort: max · single runIndependent result
LiveBench Data Analysis ↗2.0 · 2026-10-14 (dataset 2026-09-02)63.5% ↗Effort: max · single runIndependent result72.3% ↗Effort: high · single runIndependent result66.4% ↗Effort: max · single runIndependent result
LiveBench Reasoning ↗2.0 · 2026-10-14 (dataset 2026-09-02)83.8% ↗Effort: max · single runIndependent result79.6% ↗Effort: high · single runIndependent result80.2% ↗Effort: max · single runIndependent result
Specifications
Context window1M1.1M1M
Max output64K131.1K943.7K
Parameters——501B
Tool callingSupportedSupportedSupported
Structured outputSupportedSupportedSupported
Image inputSupportedSupportedNot reported
Audio inputNot reportedSupportedNot reported
Video inputSupportedSupportedNot reported
ReasoningSupportedSupportedSupported
Prompt cachingSupportedSupportedSupported
Open weightsSupportedSupportedSupported

Add another model

Search all models →
deepseek-aiDeepSeek-V4.1-Flash-1017metaMeta: Muse Spark OpenNVIDIANemotron 3.5 Super VLgoogleGoogle: Gemini 4 ArgonopenaiOpenAI: GPT-6.1 SolanthropicAnthropic: Claude Sonnet 5.5anthropicAnthropic: Claude Opus 5.5openaiOpenAI: GPT-6 Luna