Beam 1.0 is the stronger all-rounder.DeepSeek-V4-Flash is better at reasoning.
Scores updated · 17 tests both models report · How we compare
Where each one wins
Tests won in each of the five areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software41Beam 1.04 of 5 tests
- AgentsCarrying out multi-step tasks on its own20Beam 1.02 of 2 tests
- MathCompetition and research-level math10Beam 1.01 of 1 test
- ReasoningHard problems that need careful thinking02DeepSeek-V4-Flash2 of 4 tests · 2 ties
- Following instructionsDoing exactly what it is asked00Even0 each · 1 tie
Math and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Beam 1.0 pulls ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+13.5points ahead
- Multi-step tasks using many real software toolsMCP Atlas+11.8points ahead
- Fixes real GitHub issues in many programming languagesSWE-bench Multilingual+5points ahead
Where DeepSeek-V4-Flash pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+37.9points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+2.6points ahead
- Code for real scientific research problemsSciCode+1.1points ahead
Every test, side by side
All 17 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingBeam 1.0
- SWE-bench ProBeam 1.0 by 13.566.152.6+13.5
- SWE-bench MultilingualBeam 1.0 by 578.373.3+5
- SWE-bench VerifiedBeam 1.0 by 2.681.679+2.6
- Terminal-Bench 2.1Beam 1.0 by 1.480.178.7+1.4
- SciCodeDeepSeek-V4-Flash by 1.149.250.3+1.1
AgentsBeam 1.0
- MCP AtlasBeam 1.0 by 11.880.869+11.8
- BrowseCompBeam 1.0 by 3.676.873.2+3.6
ReasoningDeepSeek-V4-Flash
- ARC-AGI-2DeepSeek-V4-Flash by 37.923.561.4+37.9
- Humanity's Last ExamDeepSeek-V4-Flash by 2.63638.6+2.6
- CritPttie16.316.6tie
- GPQA Diamondtie90.690.8tie
Other results4 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- ARC-AGI-1DeepSeek-V4-Flash by 157489+15
- τ³-Bench BankingBeam 1.0 by 13.236.122.9+13.2
- DeepSWE 1.1DeepSeek-V4-Flash by 11.143.354.4+11.1
- HLE (with tools)Beam 1.0 by 651.145.1+6
Questions people ask
Which is better, Beam 1.0 or DeepSeek-V4-Flash?
Beam 1.0 wins three of the five areas where both have results: coding, agents and math. DeepSeek-V4-Flash wins reasoning. They are level on following instructions.
Which is better for coding?
Beam 1.0. It wins 4 of the 5 coding tests both models report; DeepSeek-V4-Flash wins 1.
How do you compare the two?
We use the 17 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 4 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.