Side by side
See the difference.
Scores, capabilities and costs for the models you’re considering.
Scroll to compare models →
| 3 models | ByteDanceSeed-1.8Remove × | googleembeddinggemma-2Remove × | reflectionReflection: Beam 1.0 501B A23BRemove × |
|---|---|---|---|
| What’s different | Optimized specifically for multimodal agent scenarios. It features enhanced agent capabilities, upgraded multimodal comprehension, and more flexible context management. | EmbeddingGemma 2 is a 740M parameter multimodal embedding model from Google DeepMind that maps text (incl. code), images, video and audio, alone or interleaved, into a single 768-dimensional vector space,… | Apache-2.0 text model with 501B total and 23B active parameters; ships BF16, FP8 and NVFP4 weights. |
| Date | Feb 17, 2026Listed | Oct 8, 2026Listed | Oct 16, 2026Released |
API prices USD · provider configuration shown below | |||
| Listed rate | $0.25 in / $2.00 out/ 1M tokensDeepInfra · all routes ↗ | See provider quotes | $0.25 in / $1.00 out/ 1M tokensNebius · via OpenRouter · all routes ↗nebius/fp8 · quantization=fp8 |
| Provider endpoints | 1 endpoint → | 0 endpoints → | 5 endpoints → |
Benchmarks Each result retains its evaluation configuration | |||
| AA Intelligence ↗4.3.2 | Not reported | Not reported | 30.33 ↗Effort not specifiedIndependent result |
| AA-Briefcase ↗1.1 | Not reported | Not reported | 976.2 ↗Effort not specifiedIndependent result |
| GDPval-AA ↗2.1 | Not reported | Not reported | 1,177.08 ↗Effort not specifiedIndependent result |
| AutomationBench-AA ↗AA Index 4.3.2 collection | Not reported | Not reported | 48.46% ↗Effort not specifiedIndependent result |
| Terminal-Bench 4.0 · AA ↗ | Not reported | Not reported | 3.03% ↗Effort not specifiedHarness unreportedIndependent result |
| SciCode ↗AA Index 4.3.2 collection | Not reported | Not reported | 48.5% ↗AA · Under reviewEffort not specifiedIndependent result |
| Humanity’s Last Exam ↗AA Index 4.3.2 collection | Not reported | Not reported | 34.66% ↗Effort not specifiedIndependent result |
| GDP.pdf ↗AA Index 4.3.2 collection | Not reported | Not reported | 11.6% ↗Effort not specifiedIndependent result |
| CritPt ↗AA Index 4.3.2 collection | Not reported | Not reported | 13.14% ↗AA · Under reviewEffort not specifiedIndependent result |
| AA-Omniscience ↗AA Index 4.3.2 collection | Not reported | Not reported | 1.5 ↗Effort not specifiedIndependent result |
| AA-LCR ↗1.1 | Not reported | Not reported | 78% ↗Effort not specifiedIndependent result |
| LiveBench ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 55.5% ↗Effort: max · single run · equal mean of 7 suitesIndependent result |
| LiveBench Design ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 67.9% ↗Effort: max · single runIndependent result |
| LiveBench Research ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 72.8% ↗Effort: max · single runIndependent result |
| LiveBench Browser Use ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 0% ↗Effort: max · single runIndependent result |
| LiveBench Coding ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 83.9% ↗Effort: max · single runIndependent result |
| LiveBench Agentic ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 17.3% ↗Effort: max · single runIndependent result |
| LiveBench Data Analysis ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 66.4% ↗Effort: max · single runIndependent result |
| LiveBench Reasoning ↗2.0 · 2026-10-14 (dataset 2026-09-02) | Not reported | Not reported | 80.2% ↗Effort: max · single runIndependent result |
Specifications | |||
| Context window | 256K | 8.2K | 1M |
| Max output | — | — | 943.7K |
| Parameters | — | — | 501B |
| Tool calling | Not reported | Not reported | Supported |
| Structured output | Not reported | Not reported | Supported |
| Reasoning | Not reported | Not reported | Supported |
| Prompt caching | Not reported | Not reported | Supported |
| Open weights | Not reported | Not reported | Supported |