Side by side

See the difference.

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

Scroll to compare models →

Comparison of Gemini 3.7 Flash, Bria Guided Remove Background, Reflection: Beam 1.0 501B A23B
3 modelsGoogleGemini 3.7 FlashRemove ×briaBria Guided Remove BackgroundRemove ×reflectionReflection: Beam 1.0 501B A23BRemove ×
What’s differentReasoningName the objects to keep or remove; returns the original pixels with an alpha channel, without repainting the subject.Apache-2.0 text model with 501B total and 23B active parameters; ships BF16, FP8 and NVFP4 weights.
DateOct 16, 2026DiscoveredOct 10, 2026ListedOct 16, 2026Released
Listed rateSee provider quotes$0.03/ imagefal · all routes ↗$0.25 in / $1.00 out/ 1M tokensNebius · via OpenRouter · all routes ↗nebius/fp8 · quantization=fp8
Provider endpoints0 endpoints →1 endpoint →6 endpoints →
AA Intelligence ↗4.3.2Not reportedNot reported30.33 ↗Effort not specifiedIndependent result
AA-Briefcase ↗1.1Not reportedNot reported976.2 ↗Effort not specifiedIndependent result
GDPval-AA ↗2.1Not reportedNot reported1,177.08 ↗Effort not specifiedIndependent result
AutomationBench-AA ↗AA Index 4.3.2 collectionNot reportedNot reported48.46% ↗Effort not specifiedIndependent result
Terminal-Bench 4.0 · AA ↗Not reportedNot reported3.03% ↗Effort not specifiedHarness unreportedIndependent result
SciCode ↗AA Index 4.3.2 collectionNot reportedNot reported48.5% ↗AA · Under reviewEffort not specifiedIndependent result
Humanity’s Last Exam ↗AA Index 4.3.2 collectionNot reportedNot reported34.66% ↗Effort not specifiedIndependent result
GDP.pdf ↗AA Index 4.3.2 collectionNot reportedNot reported11.6% ↗Effort not specifiedIndependent result
CritPt ↗AA Index 4.3.2 collectionNot reportedNot reported13.14% ↗AA · Under reviewEffort not specifiedIndependent result
AA-Omniscience ↗AA Index 4.3.2 collectionNot reportedNot reported1.5 ↗Effort not specifiedIndependent result
AA-LCR ↗1.1Not reportedNot reported78% ↗Effort not specifiedIndependent result
LiveBench ↗2.0 · 2026-10-14 (dataset 2026-09-02)67.19% ↗Effort: high · single run · equal mean of 7 suitesIndependent resultNot reported55.5% ↗Effort: max · single run · equal mean of 7 suitesIndependent result
LiveBench Design ↗2.0 · 2026-10-14 (dataset 2026-09-02)73.2% ↗Effort: high · single runIndependent resultNot reported67.9% ↗Effort: max · single runIndependent result
LiveBench Research ↗2.0 · 2026-10-14 (dataset 2026-09-02)76.5% ↗Effort: high · single runIndependent resultNot reported72.8% ↗Effort: max · single runIndependent result
LiveBench Browser Use ↗2.0 · 2026-10-14 (dataset 2026-09-02)56% ↗Effort: high · single runIndependent resultNot reported0% ↗Effort: max · single runIndependent result
LiveBench Coding ↗2.0 · 2026-10-14 (dataset 2026-09-02)80.5% ↗Effort: high · single runIndependent resultNot reported83.9% ↗Effort: max · single runIndependent result
LiveBench Agentic ↗2.0 · 2026-10-14 (dataset 2026-09-02)28.3% ↗Effort: high · single runIndependent resultNot reported17.3% ↗Effort: max · single runIndependent result
LiveBench Data Analysis ↗2.0 · 2026-10-14 (dataset 2026-09-02)68% ↗Effort: high · single runIndependent resultNot reported66.4% ↗Effort: max · single runIndependent result
LiveBench Reasoning ↗2.0 · 2026-10-14 (dataset 2026-09-02)87.8% ↗Effort: high · single runIndependent resultNot reported80.2% ↗Effort: max · single runIndependent result
Context window——1M
Max output——943.7K
Parameters——501B
Tool callingNot reportedNot reportedSupported
Structured outputNot reportedNot reportedSupported
ReasoningSupportedNot reportedSupported
Prompt cachingNot reportedNot reportedSupported
Open weightsNot reportedNot reportedSupported

Add another model

Search all models →